chore: generate

refactor(console): stream zen bodies without modifiers (#44472 )
fix(console): revert scanned zen request streaming (#44470 )
2026-08-23 22:23:18 -04:00 · 2026-08-23 17:15:36 +00:00 · 2026-08-23 13:14:18 -04:00 · 2026-08-23 13:03:20 -04:00 · 2026-08-23 16:51:36 +00:00 · 2026-08-23 12:50:18 -04:00
125 changed files with 3066 additions and 2204 deletions
@@ -29,7 +29,7 @@
    },
    "packages/app": {
      "name": "@opencode-ai/app",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@corvu/drawer": "catalog:",
        "@dnd-kit/abstract": "0.5.0",
@@ -96,7 +96,7 @@
    },
    "packages/cli": {
      "name": "@opencode-ai/cli",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "bin": {
        "lildax": "./bin/lildax.cjs",
      },
@@ -144,7 +144,7 @@
    },
    "packages/codemode": {
      "name": "@opencode-ai/codemode",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "acorn": "8.15.0",
        "effect": "catalog:",
@@ -158,7 +158,7 @@
    },
    "packages/console/app": {
      "name": "@opencode-ai/console-app",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@cloudflare/vite-plugin": "1.15.2",
        "@ibm/plex": "6.4.1",
@@ -194,7 +194,7 @@
    },
    "packages/console/core": {
      "name": "@opencode-ai/console-core",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@aws-sdk/client-sts": "3.782.0",
        "@jsx-email/render": "1.1.1",
@@ -221,7 +221,7 @@
    },
    "packages/console/function": {
      "name": "@opencode-ai/console-function",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@ai-sdk/anthropic": "3.0.82",
        "@ai-sdk/openai": "3.0.48",
@@ -243,7 +243,7 @@
    },
    "packages/console/mail": {
      "name": "@opencode-ai/console-mail",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@jsx-email/all": "2.2.3",
        "@jsx-email/cli": "1.4.3",
@@ -267,7 +267,7 @@
    },
    "packages/console/support": {
      "name": "@opencode-ai/console-support",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@cloudflare/vite-plugin": "1.15.2",
        "@opencode-ai/console-core": "workspace:*",
@@ -287,13 +287,13 @@
    },
    "packages/core": {
      "name": "@opencode-ai/core",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "bin": {
        "opencode": "./bin/opencode",
      },
      "dependencies": {
        "@ai-sdk/alibaba": "1.0.17",
-        "@ai-sdk/amazon-bedrock": "4.0.112",
+        "@ai-sdk/amazon-bedrock": "4.0.158",
        "@ai-sdk/anthropic": "3.0.82",
        "@ai-sdk/azure": "3.0.88",
        "@ai-sdk/cerebras": "2.0.41",
@@ -381,7 +381,7 @@
    },
    "packages/desktop": {
      "name": "@opencode-ai/desktop",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@zip.js/zip.js": "2.7.62",
        "drizzle-orm": "catalog:",
@@ -435,7 +435,7 @@
    },
    "packages/effect-drizzle-sqlite": {
      "name": "@opencode-ai/effect-drizzle-sqlite",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "drizzle-orm": "catalog:",
        "effect": "catalog:",
@@ -449,7 +449,7 @@
    },
    "packages/effect-sqlite-node": {
      "name": "@opencode-ai/effect-sqlite-node",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "effect": "catalog:",
      },
@@ -461,7 +461,7 @@
    },
    "packages/enterprise": {
      "name": "@opencode-ai/enterprise",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@hono/standard-validator": "catalog:",
        "@opencode-ai/core": "workspace:*",
@@ -493,7 +493,7 @@
    },
    "packages/function": {
      "name": "@opencode-ai/function",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@octokit/auth-app": "8.0.1",
        "@octokit/rest": "catalog:",
@@ -509,7 +509,7 @@
    },
    "packages/http-recorder": {
      "name": "@opencode-ai/http-recorder",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@effect/platform-node": "4.0.0-beta.83",
        "@effect/platform-node-shared": "4.0.0-beta.83",
@@ -540,7 +540,7 @@
    },
    "packages/llm": {
      "name": "@opencode-ai/llm",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@opencode-ai/schema": "workspace:*",
        "@smithy/eventstream-codec": "4.2.14",
@@ -559,7 +559,7 @@
    },
    "packages/opencode": {
      "name": "opencode",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "bin": {
        "opencode": "./bin/opencode",
      },
@@ -568,7 +568,7 @@
        "@actions/github": "6.0.1",
        "@agentclientprotocol/sdk": "0.21.0",
        "@ai-sdk/alibaba": "1.0.17",
-        "@ai-sdk/amazon-bedrock": "4.0.112",
+        "@ai-sdk/amazon-bedrock": "4.0.158",
        "@ai-sdk/anthropic": "3.0.82",
        "@ai-sdk/azure": "3.0.88",
        "@ai-sdk/cerebras": "2.0.60",
@@ -690,7 +690,7 @@
    },
    "packages/plugin": {
      "name": "@opencode-ai/plugin",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@ai-sdk/provider": "3.0.8",
        "@opencode-ai/sdk": "workspace:*",
@@ -766,7 +766,7 @@
    },
    "packages/sdk/js": {
      "name": "@opencode-ai/sdk",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "cross-spawn": "catalog:",
      },
@@ -781,7 +781,7 @@
    },
    "packages/server": {
      "name": "@opencode-ai/server",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@opencode-ai/core": "workspace:*",
        "@opencode-ai/protocol": "workspace:*",
@@ -796,7 +796,7 @@
    },
    "packages/session-ui": {
      "name": "@opencode-ai/session-ui",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@kobalte/core": "catalog:",
        "@opencode-ai/client": "file:../app/vendor/opencode-ai-client-1.17.13-v2.tgz",
@@ -836,7 +836,7 @@
    },
    "packages/slack": {
      "name": "@opencode-ai/slack",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@opencode-ai/sdk": "workspace:*",
        "@slack/bolt": "^3.17.1",
@@ -849,7 +849,7 @@
    },
    "packages/stats/app": {
      "name": "@opencode-ai/stats-app",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@ibm/plex": "6.4.1",
        "@kobalte/core": "catalog:",
@@ -883,7 +883,7 @@
    },
    "packages/stats/core": {
      "name": "@opencode-ai/stats-core",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@aws-sdk/client-athena": "3.933.0",
        "@planetscale/database": "1.19.0",
@@ -902,7 +902,7 @@
    },
    "packages/stats/server": {
      "name": "@opencode-ai/stats-server",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@aws-sdk/client-firehose": "3.933.0",
        "@effect/platform-node": "catalog:",
@@ -944,7 +944,7 @@
    },
    "packages/tui": {
      "name": "@opencode-ai/tui",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@opencode-ai/core": "workspace:*",
        "@opencode-ai/plugin": "workspace:*",
@@ -971,7 +971,7 @@
    },
    "packages/ui": {
      "name": "@opencode-ai/ui",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@kobalte/core": "catalog:",
        "@pierre/diffs": "catalog:",
@@ -1022,7 +1022,7 @@
    },
    "packages/web": {
      "name": "@opencode-ai/web",
-      "version": "1.18.19",
+      "version": "1.18.21",
      "dependencies": {
        "@astrojs/cloudflare": "12.6.3",
        "@astrojs/markdown-remark": "6.3.1",
@@ -1173,7 +1173,7 @@

    "@ai-sdk/alibaba": ["@ai-sdk/alibaba@1.0.17", "", { "dependencies": { "@ai-sdk/openai-compatible": "2.0.41", "@ai-sdk/provider": "3.0.8", "@ai-sdk/provider-utils": "4.0.23" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-ZbE+U5bWz2JBc5DERLowx5+TKbjGBE93LqKZAWvuEn7HOSQMraxFMZuc0ST335QZJAyfBOzh7m1mPQ+y7EaaoA=="],

-    "@ai-sdk/amazon-bedrock": ["@ai-sdk/amazon-bedrock@4.0.112", "", { "dependencies": { "@ai-sdk/anthropic": "3.0.81", "@ai-sdk/openai": "3.0.67", "@ai-sdk/provider": "3.0.10", "@ai-sdk/provider-utils": "4.0.27", "@smithy/eventstream-codec": "^4.0.1", "@smithy/util-utf8": "^4.0.0", "aws4fetch": "^1.0.20" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-PsSh7a6qW+3kQXPs1kD4wDwuZby0t1PIaB6j/1aMKmPFJ5LxcIcULLMF/bjITLt5o/8lc0t6TXIwG0zlhH7uZw=="],
+    "@ai-sdk/amazon-bedrock": ["@ai-sdk/amazon-bedrock@4.0.158", "", { "dependencies": { "@ai-sdk/anthropic": "3.0.111", "@ai-sdk/openai": "3.0.98", "@ai-sdk/provider": "3.0.15", "@ai-sdk/provider-utils": "4.0.46", "@smithy/eventstream-codec": "^4.0.1", "@smithy/util-utf8": "^4.0.0", "aws4fetch": "^1.0.20" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-yZebHEszUzPLsK+Rq5sVZJkJj7EYDgY+Lz36IGf/RSkSC5LOMDbVKoQ55S8xNzJqVjUDqlWuZiQzg40HQslmCw=="],

    "@ai-sdk/anthropic": ["@ai-sdk/anthropic@3.0.82", "", { "dependencies": { "@ai-sdk/provider": "3.0.10", "@ai-sdk/provider-utils": "4.0.27" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-WKKou2wbhGGYV8PSALAPyV2YY4nfCqCPkyBzYtJtDA9yCcIFwsbtkTNgg7bqtLCVzeEsY7wwxRoCWy+EMfrw/A=="],

@@ -5641,13 +5641,13 @@

    "@ai-sdk/alibaba/@ai-sdk/openai-compatible": ["@ai-sdk/openai-compatible@2.0.41", "", { "dependencies": { "@ai-sdk/provider": "3.0.8", "@ai-sdk/provider-utils": "4.0.23" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-kNAGINk71AlOXx10Dq/PXw4t/9XjdK8uxfpVElRwtSFMdeSiLVt58p9TPx4/FJD+hxZuVhvxYj9r42osxWq79g=="],

-    "@ai-sdk/amazon-bedrock/@ai-sdk/anthropic": ["@ai-sdk/anthropic@3.0.81", "", { "dependencies": { "@ai-sdk/provider": "3.0.10", "@ai-sdk/provider-utils": "4.0.27" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-B1JDd9Ugq9R5AgIaW3674lhGCMMYJcPUxnrZh8fzbGojgg4QvHFRv6eZahGQAUsmGHbcf74G9bdSBDLWQGY2GA=="],
+    "@ai-sdk/amazon-bedrock/@ai-sdk/anthropic": ["@ai-sdk/anthropic@3.0.111", "", { "dependencies": { "@ai-sdk/provider": "3.0.15", "@ai-sdk/provider-utils": "4.0.46" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-atgBW8jZPr/KuaKX5FvDIHuXBI8VCol6kVeoD4P0657+VXR73QsLogXQVN/Zt5FHtq9WzpdIZseCJiXPqkgwwA=="],

-    "@ai-sdk/amazon-bedrock/@ai-sdk/openai": ["@ai-sdk/openai@3.0.67", "", { "dependencies": { "@ai-sdk/provider": "3.0.10", "@ai-sdk/provider-utils": "4.0.27" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-oAiGC9eWG7IgtdsdS74bOCnAAHarAfTJhWN9x5INwnWPekL802AvF+0I5DvLzIF1MIRmNw4N8mPSL/GUVbX9Mw=="],
+    "@ai-sdk/amazon-bedrock/@ai-sdk/openai": ["@ai-sdk/openai@3.0.98", "", { "dependencies": { "@ai-sdk/provider": "3.0.15", "@ai-sdk/provider-utils": "4.0.46" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-nAvp8pVOUJ3znJHRzZs54Y7CJQSikOW1Ty7LTdtQT+/pgtAwqhuKd8oaXbiW2xQaYBdH2i8o9AcEpcUdXIyF+g=="],

-    "@ai-sdk/amazon-bedrock/@ai-sdk/provider": ["@ai-sdk/provider@3.0.10", "", { "dependencies": { "json-schema": "^0.4.0" } }, "sha512-Q3BZ27qfpYqnCYGvE3vt+Qi6LGOF9R5Nmzn+9JoM1lCRsD9mYaIhfJLkSunN48nfGXJ6n+XNV0J/XVpqGQl7Dw=="],
+    "@ai-sdk/amazon-bedrock/@ai-sdk/provider": ["@ai-sdk/provider@3.0.15", "", { "dependencies": { "json-schema": "^0.4.0" } }, "sha512-XeZW1CcDF2GMbH4wejW6xBRI2QCOgnkVYUnxoeDadB1mf85riL2bMUeDoh+6gJ/r4mjNfzUPW8OjLjvwTP0u1Q=="],

-    "@ai-sdk/amazon-bedrock/@ai-sdk/provider-utils": ["@ai-sdk/provider-utils@4.0.27", "", { "dependencies": { "@ai-sdk/provider": "3.0.10", "@standard-schema/spec": "^1.1.0", "eventsource-parser": "^3.0.8" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-ubkAJ+xODouwtmN1tYlvTPphH1hPOBfZaEQe8U7skGvFAnIRs9PPpsq57bC2+Ky/MB4yzhd6YOsxTAx9sGpazw=="],
+    "@ai-sdk/amazon-bedrock/@ai-sdk/provider-utils": ["@ai-sdk/provider-utils@4.0.46", "", { "dependencies": { "@ai-sdk/provider": "3.0.15", "@standard-schema/spec": "^1.1.0", "eventsource-parser": "^3.0.8", "undici": "^6.28.0" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-tEtld97plCFiYevsJuOkGkeuhQndeMWFBVrJS4AjnbD5AqrNSXRCe0p+BZ3Cju/sxDeeZ9ym3q9YUV8fASA7aQ=="],

    "@ai-sdk/amazon-bedrock/@smithy/eventstream-codec": ["@smithy/eventstream-codec@4.2.14", "", { "dependencies": { "@aws-crypto/crc32": "5.2.0", "@smithy/types": "^4.14.1", "@smithy/util-hex-encoding": "^4.2.2", "tslib": "^2.6.2" } }, "sha512-erZq0nOIpzfeZdCyzZjdJb4nVSKLUmSkaQUVkRGQTXs30gyUGeKnrYEg+Xe1W5gE3aReS7IgsvANwVPxSzY6Pw=="],

@@ -6589,6 +6589,8 @@

    "@ai-sdk/amazon-bedrock/@ai-sdk/provider-utils/@standard-schema/spec": ["@standard-schema/spec@1.1.0", "", {}, "sha512-l2aFy5jALhniG5HgqrD6jXLi/rUWrKvqN/qJx6yoJsgKhblVd+iqqU4RCXavm/jPityDo5TCvKMnpjKnOriy0w=="],

+    "@ai-sdk/amazon-bedrock/@ai-sdk/provider-utils/undici": ["undici@6.28.0", "", {}, "sha512-LIY910g9TI13YS95lrMFrs8Rm/u/irgHeTWoKCoteeJ04CUJ92eEfj0rVn+7VKMPBpUPiUoBKfhNyLI23EE/KA=="],
+
    "@ai-sdk/anthropic/@ai-sdk/provider-utils/@standard-schema/spec": ["@standard-schema/spec@1.1.0", "", {}, "sha512-l2aFy5jALhniG5HgqrD6jXLi/rUWrKvqN/qJx6yoJsgKhblVd+iqqU4RCXavm/jPityDo5TCvKMnpjKnOriy0w=="],

    "@ai-sdk/azure/@ai-sdk/provider-utils/@standard-schema/spec": ["@standard-schema/spec@1.1.0", "", {}, "sha512-l2aFy5jALhniG5HgqrD6jXLi/rUWrKvqN/qJx6yoJsgKhblVd+iqqU4RCXavm/jPityDo5TCvKMnpjKnOriy0w=="],
@@ -1,8 +1,8 @@
 {
  "nodeModules": {
-    "x86_64-linux": "sha256-yQ8EIxxYkzlEWIMY/UiIR+7lbGBuizsoehMPZzFA65Y=",
-    "aarch64-linux": "sha256-JF9VVgnl5QUZ430fqb5Qu8y0kchYJ00LO3FbYcd0lBM=",
-    "aarch64-darwin": "sha256-f3Tu6eu463NWcHgr7dupfD/zUTh26bJ6N2vVXuEyi6c=",
-    "x86_64-darwin": "sha256-miv9Sv4KdhD0UIi2O5LVS1wQOfq29IV+9S+BvIzAvgo="
+    "x86_64-linux": "sha256-Be1I6OG6UitofhcGu2BeNzevmoQXc4Or5r/NPzwtft4=",
+    "aarch64-linux": "sha256-O+d+26CQIjZ08Rn8Qm3IytdDqIMbPdhzaGOZeUKAvIU=",
+    "aarch64-darwin": "sha256-ObS50y/oy6fM9wSGUL/wx6O0+fTWHC04mXJNd7w/2Z0=",
+    "x86_64-darwin": "sha256-eoR7ZSyH62Fq2ZaW2b2QqU2FC97rYxTMeEe+djT0nto="
  }
 }
@@ -60,6 +60,39 @@ test("keeps the file-browser sidebar mounted when switching file tabs", async ({
  await expect.poll(() => viewport.evaluate((element) => element.scrollTop)).toBe(scrolled)
 })

+test("keeps previous file search results visible while the next search loads", async ({ page }) => {
+  const searchPending = Promise.withResolvers<void>()
+  await setup(page, async ({ query }) => {
+    if (query === "file-0") return ["file-00.ts"]
+    if (query === "file-7") {
+      await searchPending.promise
+      return ["file-79.ts"]
+    }
+    return []
+  })
+
+  await page.goto(`/server/${base64Encode(server)}/session/${sessionID}`)
+  await expectSessionTitle(page, title)
+
+  const panel = page.locator("#review-panel")
+  await panel.getByRole("button", { name: "Open file" }).click()
+  const filter = panel.getByRole("combobox", { name: "Filter files" })
+  await filter.fill("file-0")
+  await expect(panel.getByRole("option", { name: "file-00.ts" })).toBeVisible()
+
+  const nextSearch = page.waitForRequest((request) => {
+    const url = new URL(request.url())
+    return url.pathname === "/find/file" && url.searchParams.get("query") === "file-7"
+  })
+  await filter.fill("file-7")
+  await nextSearch
+  await expect(panel.getByRole("option", { name: "file-00.ts" })).toBeVisible()
+
+  searchPending.resolve()
+  await expect(panel.getByRole("option", { name: "file-79.ts" })).toBeVisible()
+  await expect(panel.getByRole("option", { name: "file-00.ts" })).toBeHidden()
+})
+
 type Probed = HTMLElement & { __e2eProbe?: string }

 async function writeProbe(page: Page) {
@@ -74,7 +107,10 @@ async function readProbe(page: Page) {
    .evaluate((el) => (el as Probed).__e2eProbe)
 }

-async function setup(page: Page) {
+async function setup(
+  page: Page,
+  findFiles?: (input: { query: string; dirs?: string; limit?: number }) => unknown | Promise<unknown>,
+) {
  await mockOpenCodeServer(page, {
    directory,
    project: {
@@ -119,6 +155,7 @@ async function setup(page: Page) {
      }))
    },
    fileContent: (path) => ({ type: "text", content: `contents:${path}` }),
+    findFiles,
    pageMessages: () => ({ items: [] }),
  })

@@ -26,7 +26,7 @@ export interface MockServerConfig {
  questions?: unknown[] | (() => unknown[])
  fileList?: (path: string) => unknown | Promise<unknown>
  fileContent?: (path: string) => unknown | Promise<unknown>
-  findFiles?: (input: { query: string; dirs?: string; limit?: number }) => unknown
+  findFiles?: (input: { query: string; dirs?: string; limit?: number }) => unknown | Promise<unknown>
  sessionStatus?: Record<string, unknown> | (() => Record<string, unknown>)
 }

@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/app",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "description": "",
  "type": "module",
  "exports": {
@@ -450,7 +450,7 @@ function ModelSelectorPopoverV2View(props: {
                <For each={groups()}>
                  {(group) => (
                    <MenuV2.Group>
-                      <MenuV2.GroupLabel class="gap-2 px-3">
+                      <MenuV2.GroupLabel class="sticky top-0 z-10 gap-2 bg-v2-background-bg-layer-01 px-3">
                        <span class="min-w-0 truncate">{group.items[0].provider.name}</span>
                      </MenuV2.GroupLabel>
                      <MenuV2.RadioGroup value={props.current()}>
@@ -972,17 +972,6 @@ export default function LegacyLayout(props: ParentProps) {
        keybind: "shift+alt+arrowdown",
        onSelect: () => navigateSessionByUnseen(1),
      },
-      {
-        id: "session.archive",
-        title: language.t("command.session.archive"),
-        category: language.t("command.category.session"),
-        keybind: "mod+shift+backspace",
-        disabled: !params.dir || !params.id,
-        onSelect: () => {
-          const session = currentSessions().find((s) => s.id === params.id)
-          if (session) void archiveSession(session)
-        },
-      },
      {
        id: "workspace.new",
        title: language.t("workspace.new"),
@@ -0,0 +1,71 @@
+import { useNavigate } from "@solidjs/router"
+import { produce } from "solid-js/store"
+import { notifySessionTabsRemoved } from "@/components/titlebar-session-events"
+import { useLanguage } from "@/context/language"
+import { useSDK } from "@/context/sdk"
+import { useSync } from "@/context/sync"
+import { useTabs } from "@/context/tabs"
+import { errorMessage } from "@/pages/layout/helpers"
+import { useSessionKey } from "@/pages/session/session-layout"
+import { legacySessionHref, requireServerKey, sessionHref } from "@/utils/session-route"
+import { showToast } from "@/utils/toast"
+
+export function useSessionArchive() {
+  const language = useLanguage()
+  const navigate = useNavigate()
+  const sdk = useSDK()
+  const sync = useSync()
+  const tabs = useTabs()
+  const { params } = useSessionKey()
+
+  const navigateAfterRemoval = (sessionID: string, parentID?: string, nextSessionID?: string) => {
+    if (params.id !== sessionID) return
+    const href = (id: string) =>
+      params.serverKey ? sessionHref(requireServerKey(params.serverKey), id) : legacySessionHref(sdk().directory, id)
+    if (parentID) {
+      navigate(href(parentID))
+      return
+    }
+    if (nextSessionID) {
+      navigate(href(nextSessionID))
+      return
+    }
+    if (params.serverKey) {
+      tabs.newDraft({ server: requireServerKey(params.serverKey), directory: sdk().directory })
+      return
+    }
+    navigate(`/${params.dir}/session`)
+  }
+
+  const archive = async (sessionID: string) => {
+    const session = sync().session.get(sessionID)
+    if (!session) return
+    if ((await sdk().protocol) !== "v1") return
+
+    const sessions = sync().data.session ?? []
+    const index = sessions.findIndex((s) => s.id === sessionID)
+    const nextSession = index === -1 ? undefined : (sessions[index + 1] ?? sessions[index - 1])
+
+    await sdk()
+      .client.session.update({ sessionID, directory: sdk().directory, time: { archived: Date.now() } })
+      .then(() => {
+        sync().set(
+          produce((draft) => {
+            const index = draft.session.findIndex((s) => s.id === sessionID)
+            if (index !== -1) draft.session.splice(index, 1)
+          }),
+        )
+        sync().session.evict(sessionID)
+        navigateAfterRemoval(sessionID, session.parentID, nextSession?.id)
+        notifySessionTabsRemoved({ directory: sdk().directory, sessionIDs: [sessionID] })
+      })
+      .catch((err) => {
+        showToast({
+          title: language.t("common.requestFailed"),
+          description: errorMessage(err, language.t("common.requestFailed")),
+        })
+      })
+  }
+
+  return { archive, navigateAfterRemoval }
+}
@@ -63,10 +63,10 @@ import { SessionContextUsage } from "@/components/session-context-usage"
 import { useDialog } from "@opencode-ai/ui/context/dialog"
 import { useLanguage } from "@/context/language"
 import { useSessionKey } from "@/pages/session/session-layout"
+import { useSessionArchive } from "@/pages/session/session-archive"
 import { useServerSDK } from "@/context/server-sdk"
 import { usePlatform } from "@/context/platform"
 import { useSettings } from "@/context/settings"
-import { useTabs } from "@/context/tabs"
 import { legacySessionHref, requireServerKey, sessionHref } from "@/utils/session-route"
 import { useSDK } from "@/context/sdk"
 import { useSync } from "@/context/sync"
@@ -263,8 +263,8 @@ export function MessageTimeline(props: {
  const sdk = useSDK()
  const sync = useSync()
  const settings = useSettings()
-  const tabs = useTabs()
  const dialog = useDialog()
+  const sessionArchive = useSessionArchive()
  const language = useLanguage()
  const { params, sessionKey } = useSessionKey()
  const ownerSessionKey = sessionKey()
@@ -790,25 +790,6 @@ export function MessageTimeline(props: {
    titleMutation.mutate({ id, title: next })
  }

-  const navigateAfterSessionRemoval = (sessionID: string, parentID?: string, nextSessionID?: string) => {
-    if (params.id !== sessionID) return
-    const href = (id: string) =>
-      params.serverKey ? sessionHref(requireServerKey(params.serverKey), id) : legacySessionHref(sdk().directory, id)
-    if (parentID) {
-      navigate(href(parentID))
-      return
-    }
-    if (nextSessionID) {
-      navigate(href(nextSessionID))
-      return
-    }
-    if (params.serverKey) {
-      tabs.newDraft({ server: requireServerKey(params.serverKey), directory: sdk().directory })
-      return
-    }
-    navigate(`/${params.dir}/session`)
-  }
-
  const exportSession = async (sessionID: string) => {
    try {
      const data = await fetchSessionExport({
@@ -832,36 +813,6 @@ export function MessageTimeline(props: {
    }
  }

-  const archiveSession = async (sessionID: string) => {
-    const session = sync().session.get(sessionID)
-    if (!session) return
-    if ((await sdk().protocol) !== "v1") return
-
-    const sessions = sync().data.session ?? []
-    const index = sessions.findIndex((s) => s.id === sessionID)
-    const nextSession = index === -1 ? undefined : (sessions[index + 1] ?? sessions[index - 1])
-
-    await sdk()
-      .client.session.update({ sessionID, directory: sdk().directory, time: { archived: Date.now() } })
-      .then(() => {
-        sync().set(
-          produce((draft) => {
-            const index = draft.session.findIndex((s) => s.id === sessionID)
-            if (index !== -1) draft.session.splice(index, 1)
-          }),
-        )
-        sync().session.evict(sessionID)
-        navigateAfterSessionRemoval(sessionID, session.parentID, nextSession?.id)
-        notifySessionTabsRemoved({ directory: sdk().directory, sessionIDs: [sessionID] })
-      })
-      .catch((err) => {
-        showToast({
-          title: language.t("common.requestFailed"),
-          description: errorMessage(err),
-        })
-      })
-  }
-
  const deleteSession = async (sessionID: string) => {
    const session = sync().session.get(sessionID)
    if (!session) return false
@@ -911,7 +862,7 @@ export function MessageTimeline(props: {
      }
    }

-    navigateAfterSessionRemoval(sessionID, session.parentID, nextSession?.id)
+    sessionArchive.navigateAfterRemoval(sessionID, session.parentID, nextSession?.id)

    sync().set(
      produce((draft) => {
@@ -1593,7 +1544,7 @@ export function MessageTimeline(props: {
                                <DropdownMenu.Item onSelect={() => exportSession(id)}>
                                  <DropdownMenu.ItemLabel>{language.t("common.export")}</DropdownMenu.ItemLabel>
                                </DropdownMenu.Item>
-                                <DropdownMenu.Item onSelect={() => void archiveSession(id)}>
+                                <DropdownMenu.Item onSelect={() => void sessionArchive.archive(id)}>
                                  <DropdownMenu.ItemLabel>{language.t("common.archive")}</DropdownMenu.ItemLabel>
                                </DropdownMenu.Item>
                                <DropdownMenu.Separator />
@@ -1667,7 +1618,7 @@ export function MessageTimeline(props: {
                              <MenuV2.Item onSelect={() => exportSession(id)}>
                                {language.t("common.export")}...
                              </MenuV2.Item>
-                              <MenuV2.Item onSelect={() => void archiveSession(id)}>
+                              <MenuV2.Item onSelect={() => void sessionArchive.archive(id)}>
                                {language.t("common.archive")}
                              </MenuV2.Item>
                              <MenuV2.Separator />
@@ -18,6 +18,7 @@ import { createSessionTabs } from "@/pages/session/helpers"
 import { extractPromptFromParts } from "@/utils/prompt"
 import { Message, Part, UserMessage } from "@opencode-ai/sdk/v2"
 import { useSessionLayout } from "@/pages/session/session-layout"
+import { useSessionArchive } from "@/pages/session/session-archive"
 import { createSessionOwnership } from "./session-ownership"
 import { useLocal } from "@/context/local"

@@ -52,6 +53,7 @@ export const useSessionCommands = (actions: SessionCommandContext) => {
  const navigate = useNavigate()
  const { params, sessionKey, tabs, view } = useSessionLayout()
  const sessionOwnership = createSessionOwnership(sessionKey)
+  const sessionArchive = useSessionArchive()
  const openDialog = async <T,>(load: () => Promise<T>, show: (value: T) => void) => {
    const owner = sessionOwnership.capture()
    const value = await load()
@@ -497,6 +499,16 @@ export const useSessionCommands = (actions: SessionCommandContext) => {
      disabled: !params.id,
      onSelect: exportSession,
    }),
+    sessionCommand({
+      id: "session.archive",
+      title: language.t("command.session.archive"),
+      keybind: "mod+shift+backspace",
+      disabled: !params.id,
+      onSelect: () => {
+        const id = params.id
+        if (id) void sessionArchive.archive(id)
+      },
+    }),
  ]

  const fileCmds = () => {
@@ -1,5 +1,5 @@
 import { createMemo, createSignal, createUniqueId, Show } from "solid-js"
-import { createQuery } from "@tanstack/solid-query"
+import { createQuery, keepPreviousData } from "@tanstack/solid-query"
 import { Icon } from "@opencode-ai/ui/icon"
 import { SessionFilePanelV2, SessionFilePanelV2Empty } from "@opencode-ai/session-ui/v2/session-file-panel-v2"
 import { SessionReviewV2Sidebar } from "@opencode-ai/session-ui/v2/session-review-v2"
@@ -50,6 +50,7 @@ export function SessionFileBrowserTab(props: {
      queryKey: ["session-open-file", workspaceKey(), value] as const,
      enabled: value.length > 0,
      queryFn: ({ signal }) => file.searchFiles(value, { limit: 200, signal }),
+      placeholderData: keepPreviousData,
    }
  })
  const files = createMemo(() => {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/cli",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "bin": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/codemode",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "description": "Effect-native confined code execution over schema-described tools",
  "private": true,
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-app",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -271,6 +271,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "الطلبات كل 5 ساعات",
  "go.graph.limitedRegions": "مناطق محدودة",
+  "go.graph.limitedTime": "لفترة محدودة",
  "go.graph.usageLimits": "حدود الاستخدام",
  "go.graph.aria": "الطلبات كل 5 ساعات: {{free}} مقابل {{go}}",

@@ -275,6 +275,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Requisições por 5 horas",
  "go.graph.limitedRegions": "regiões limitadas",
+  "go.graph.limitedTime": "tempo limitado",
  "go.graph.usageLimits": "Limites de uso",
  "go.graph.aria": "Requisições por 5h: {{free}} vs {{go}}",

@@ -273,6 +273,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Forespørgsler pr. 5 timer",
  "go.graph.limitedRegions": "begrænsede regioner",
+  "go.graph.limitedTime": "begrænset periode",
  "go.graph.usageLimits": "Brugsgrænser",
  "go.graph.aria": "Forespørgsler pr. 5t: {{free}} vs {{go}}",

@@ -275,6 +275,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Anfragen pro 5 Stunden",
  "go.graph.limitedRegions": "begrenzte Regionen",
+  "go.graph.limitedTime": "begrenzte Zeit",
  "go.graph.usageLimits": "Nutzungslimits",
  "go.graph.aria": "Anfragen pro 5h: {{free}} vs {{go}}",

@@ -271,6 +271,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Requests per 5 hour",
  "go.graph.limitedRegions": "limited regions",
+  "go.graph.limitedTime": "limited time",
  "go.graph.tick": "{{n}}x",
  "go.graph.usageLimits": "Usage limits",
  "go.graph.aria": "Requests per 5h: {{free}} vs {{go}}",
@@ -276,6 +276,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Solicitudes por 5 horas",
  "go.graph.limitedRegions": "regiones limitadas",
+  "go.graph.limitedTime": "tiempo limitado",
  "go.graph.usageLimits": "Límites de uso",
  "go.graph.aria": "Solicitudes por 5h: {{free}} vs {{go}}",

@@ -277,6 +277,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Requêtes par tranche de 5 heures",
  "go.graph.limitedRegions": "régions limitées",
+  "go.graph.limitedTime": "durée limitée",
  "go.graph.usageLimits": "Limites d'utilisation",
  "go.graph.aria": "Requêtes par 5h : {{free}} vs {{go}}",

@@ -273,6 +273,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Richieste ogni 5 ore",
  "go.graph.limitedRegions": "regioni limitate",
+  "go.graph.limitedTime": "periodo limitato",
  "go.graph.usageLimits": "Limiti di utilizzo",
  "go.graph.aria": "Richieste ogni 5h: {{free}} vs {{go}}",

@@ -272,6 +272,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "5時間あたりのリクエスト数",
  "go.graph.limitedRegions": "一部の地域に限定",
+  "go.graph.limitedTime": "期間限定",
  "go.graph.usageLimits": "利用制限",
  "go.graph.aria": "5時間あたりのリクエスト数: {{free}} 対 {{go}}",

@@ -269,6 +269,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "5시간당 요청 수",
  "go.graph.limitedRegions": "일부 지역에서만 제공",
+  "go.graph.limitedTime": "한정된 기간",
  "go.graph.usageLimits": "사용 한도",
  "go.graph.aria": "5시간당 요청 수: {{free}} 대 {{go}}",

@@ -273,6 +273,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Forespørsler per 5 timer",
  "go.graph.limitedRegions": "begrensede regioner",
+  "go.graph.limitedTime": "begrenset periode",
  "go.graph.usageLimits": "Bruksgrenser",
  "go.graph.aria": "Forespørsler per 5t: {{free}} vs {{go}}",

@@ -274,6 +274,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Żądania na 5 godzin",
  "go.graph.limitedRegions": "ograniczone regiony",
+  "go.graph.limitedTime": "ograniczony czas",
  "go.graph.usageLimits": "Limity użycia",
  "go.graph.aria": "Żądania na 5h: {{free}} vs {{go}}",

@@ -277,6 +277,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Запросов за 5 часов",
  "go.graph.limitedRegions": "доступно в отдельных регионах",
+  "go.graph.limitedTime": "ограниченное время",
  "go.graph.usageLimits": "Лимиты использования",
  "go.graph.aria": "Запросов за 5ч: {{free}} против {{go}}",

@@ -271,6 +271,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "คำขอต่อ 5 ชั่วโมง",
  "go.graph.limitedRegions": "เฉพาะบางภูมิภาค",
+  "go.graph.limitedTime": "ช่วงเวลาจำกัด",
  "go.graph.usageLimits": "ขีดจำกัดการใช้งาน",
  "go.graph.aria": "คำขอต่อ 5 ชม.: {{free}} vs {{go}}",

@@ -275,6 +275,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "5 saat başına istekler",
  "go.graph.limitedRegions": "sınırlı bölgeler",
+  "go.graph.limitedTime": "sınırlı süre",
  "go.graph.usageLimits": "Kullanım limitleri",
  "go.graph.aria": "5 saatlik istekler: {{free}} vs {{go}}",

@@ -274,6 +274,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "Запитів за 5 годин",
  "go.graph.limitedRegions": "обмежені регіони",
+  "go.graph.limitedTime": "обмежений час",
  "go.graph.usageLimits": "Ліміти використання",
  "go.graph.aria": "Запитів за 5 год: {{free}} vs {{go}}",

@@ -261,6 +261,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "每 5 小时请求数",
  "go.graph.limitedRegions": "仅限部分地区",
+  "go.graph.limitedTime": "限时",
  "go.graph.usageLimits": "使用限制",
  "go.graph.aria": "每 5 小时请求数: {{free}} vs {{go}}",

@@ -261,6 +261,7 @@ export const dict = {
  "go.graph.go": "Go",
  "go.graph.label": "每 5 小時請求數",
  "go.graph.limitedRegions": "僅限部分地區",
+  "go.graph.limitedTime": "限時",
  "go.graph.usageLimits": "使用限制",
  "go.graph.aria": "每 5 小時請求數：{{free}} vs {{go}}",

@@ -0,0 +1,21 @@
+import type { APIEvent } from "@solidjs/start/server"
+import { Workspace } from "@opencode-ai/console-core/workspace.js"
+import { safeEqual } from "@opencode-ai/console-core/util/crypto.js"
+import { Resource } from "@opencode-ai/console-resource"
+import z from "zod"
+
+const Body = z.object({ workspaceID: z.string().startsWith("wrk_") })
+
+export async function POST(event: APIEvent) {
+  if (!safeEqual(event.request.headers.get("authorization") ?? "", `Bearer ${Resource.SUPPORT_API_KEY.value}`)) {
+    return Response.json({ error: "Unauthorized" }, { status: 401 })
+  }
+
+  const body = Body.safeParse(await event.request.json().catch(() => undefined))
+  if (!body.success) {
+    return Response.json({ error: "Invalid request", issues: body.error.issues }, { status: 400 })
+  }
+  return Workspace.unblock(body.data)
+    .then(() => Response.json({ success: true, message: "Workspace unblocked" }))
+    .catch((error) => Response.json({ error: error instanceof Error ? error.message : String(error) }, { status: 400 }))
+}
@@ -1,4 +1,5 @@
 import type { APIEvent } from "@solidjs/start"
+import { waitUntil } from "@opencode-ai/console-resource"
 import type { DownloadPlatform } from "../types"

 const prodAssetNames: Record<string, string> = {
@@ -30,14 +31,34 @@ export async function GET({ params: { platform, channel } }: APIEvent) {
  const assetName = channel === "stable" ? prodAssetNames[platform] : betaAssetNames[platform]
  if (!assetName) return new Response(null, { status: 404 })

-  const resp = await fetch(
+  const latest = await fetch(
    `https://github.com/anomalyco/${channel === "stable" ? "opencode" : "opencode-beta"}/releases/latest/download/${assetName}`,
+    { redirect: "manual" },
  )
+  const location = latest.headers.get("location")
+  if (!location) return new Response(null, { status: 502 })

-  const downloadName = downloadNames[platform]
+  const key = new Request(location)
+  const cache = (caches as CacheStorage & { default: Cache }).default
+  const cached = await cache.match(key)
+  if (cached) return download(cached, platform, "HIT")
+
+  const resp = await fetch(location)
+  if (!resp.ok) return resp

+  const headers = new Headers(resp.headers)
+  headers.set("cache-control", "public, max-age=31536000, immutable")
+  headers.delete("set-cookie")
+  const result = new Response(resp.body, { status: resp.status, statusText: resp.statusText, headers })
+  waitUntil(cache.put(key, result.clone()))
+  return download(result, platform, "MISS")
+}
+
+function download(resp: Response, platform: string, cache: "HIT" | "MISS") {
+  const downloadName = downloadNames[platform]
  const headers = new Headers(resp.headers)
  if (downloadName) headers.set("content-disposition", `attachment; filename="${downloadName}"`)
+  headers.set("x-opencode-cache", cache)

  return new Response(resp.body, { status: resp.status, statusText: resp.statusText, headers })
 }
@@ -592,7 +592,7 @@ body {
          white-space: nowrap;
        }

-        [data-regions] {
+        :is([data-regions], [data-limited]) {
          color: var(--color-text-weak);
          white-space: nowrap;

@@ -698,6 +698,11 @@ body {
        fill: color-mix(in srgb, var(--bar-go) 50%, transparent);
      }

+      [data-item][data-infinite] [data-value] {
+        font-size: 18px;
+        line-height: 1;
+      }
+
      [data-val] {
        fill: var(--color-text-strong);
        font-size: 13px;
@@ -80,6 +80,7 @@ function LimitsGraph(props: { href: string }) {
    { id: "mimo-v2.5", name: "MiMo-V2.5", req: 30100, d: "340ms" },
    { id: "hy3", name: "Hy3", req: 34400, baseReq: 4300, d: "320ms" },
    { id: "muse-spark-1.2-contributor", name: "Muse Spark 1.2 Contributor", req: 45300, edge: true, d: "360ms" },
+    { id: "ox-alpha-free", name: "Ox Alpha Free", req: Infinity, infinite: true, edge: true, d: "400ms" },
  ]

  const w = 1040
@@ -89,9 +90,10 @@ function LimitsGraph(props: { href: string }) {
  const top = 18
  const bottom = 44
  const plot = chartW - left - right
+  const infiniteX = w - 180

  const ratio = (n: number) => n / baseline
-  const rmax = Math.max(1, ...graph.map((m) => ratio(m.req)))
+  const rmax = Math.max(1, ...graph.filter((m) => !("infinite" in m)).map((m) => ratio(m.req)))
  const log = (n: number) => Math.log10(Math.max(n, 1))
  const base = 24
  const p = 2.2
@@ -153,10 +155,10 @@ function LimitsGraph(props: { href: string }) {
                  <rect
                    x={left}
                    y={gy(i()) - bh / 2}
-                    width={Math.max(0, x(ratio(m.baseReq ?? m.req)) - left)}
+                    width={Math.max(0, ("infinite" in m ? infiniteX : x(ratio(m.baseReq ?? m.req))) - left)}
                    height={bh}
                    data-bar
-                    data-kind="go"
+                    data-kind={"infinite" in m ? "infinite" : "go"}
                    data-model={m.id}
                  />
                  {m.baseReq && (
@@ -200,9 +202,12 @@ function LimitsGraph(props: { href: string }) {
                data-kind="go"
                data-model={m.id}
                data-edge={"edge" in m ? "" : undefined}
-                style={{ "--x": px(x(ratio(m.req))), "--y": py(gy(i())), "--d": m.d } as any}
+                data-infinite={"infinite" in m ? "" : undefined}
+                style={
+                  { "--x": px("infinite" in m ? infiniteX : x(ratio(m.req))), "--y": py(gy(i())), "--d": m.d } as any
+                }
              >
-                <span data-value>{m.req.toLocaleString()}</span>
+                <span data-value>{"infinite" in m ? "∞" : m.req.toLocaleString()}</span>
                <span data-name>{m.name}</span>
                {m.id === "muse-spark-1.2-contributor" && (
                  <span data-regions>
@@ -213,6 +218,7 @@ function LimitsGraph(props: { href: string }) {
                    )
                  </span>
                )}
+                {"infinite" in m && <span data-limited>({i18n.t("go.graph.limitedTime")})</span>}
                {m.baseReq && <span data-bonus>8x usage</span>}
              </span>
            )}
@@ -356,9 +356,11 @@ export function LiteSection(props: { lite: LiteSubscription | undefined }) {
            <li>Qwen3.6 Plus</li>
            <li>DeepSeek V4 Pro</li>
            <li>DeepSeek V4 Flash</li>
+            <li>DeepSeek V4 Flash Vision Exp</li>
            <li>MiMo-V2.5</li>
            <li>MiMo-V2.5-Pro</li>
            <li>Hy3</li>
+            <li>Ox Alpha Free</li>
          </ul>
          <p data-slot="promo-description">{i18n.t("workspace.lite.promo.footer")}</p>
          <div data-slot="subscribe-actions">
@@ -28,13 +28,7 @@ import {
  GoUsageLimitError,
  BlackUsageLimitError,
 } from "./error"
-import {
-  buildCostChunk,
-  createBodyConverter,
-  createStreamPartConverter,
-  createResponseConverter,
-  UsageInfo,
-} from "./provider/provider"
+import { buildCostChunk, createStreamPartConverter, createResponseConverter, UsageInfo } from "./provider/provider"
 import { anthropicHelper } from "./provider/anthropic"
 import { googleHelper } from "./provider/google"
 import { openaiHelper } from "./provider/openai"
@@ -53,12 +47,10 @@ import { createProviderBudgetTracker } from "./providerBudgetTracker"
 import { accumulateUsage, HOT_WORKSPACES } from "./usageBatcher"
 import { Workspace } from "@opencode-ai/console-core/workspace.js"
 import { countryFromRequest, isModelCountryRestricted } from "~/lib/request-country"
+import { prepareRequestBody } from "./requestBody"

 type ZenData = Awaited<ReturnType<typeof ZenData.list>>
-type RetryOptions = {
-  excludeProviders: string[]
-  retryCount: number
-}
+type PreparedBody = Awaited<ReturnType<typeof prepareRequestBody>>
 type BillingSource = "anonymous" | "free" | "byok" | "subscription" | "lite" | "balance"

 function resolve(text: string, params?: Record<string, string | number>) {
@@ -86,8 +78,6 @@ export async function handler(
  type ProviderInfo = Awaited<ReturnType<typeof selectProvider>>
  type CostInfo = ReturnType<typeof calculateCost>

-  const MAX_FAILOVER_RETRIES = 3
-  const MAX_RETRYABLE_STATUS_RETRIES = 3
  const dict = i18n(localeFromRequest(input.request))
  const t = (key: Key, params?: Record<string, string | number>) => resolve(dict[key], params)
  const ADMIN_WORKSPACES = [
@@ -96,12 +86,14 @@ export async function handler(
    "wrk_01KKZDKDWCS1VTJF8QTX62DD50", // contributors
  ]

+  let requestBody: PreparedBody | undefined
  try {
    const url = input.request.url
-    const body = await input.request.json()
-    const model = opts.parseModel(url, body)
-    const variant = opts.parseVariant(url, body)
-    const isStream = opts.parseIsStream(url, body)
+    const body = input.request.body
+    if (!body) throw new Error("Missing request body")
+    requestBody = opts.format === "google" ? undefined : await prepareRequestBody(body)
+    const model = opts.format === "google" ? opts.parseModel(url, undefined) : (requestBody?.model ?? "")
+    const googleStream = opts.format === "google" ? opts.parseIsStream(url, undefined) : undefined
    const rawIp = input.request.headers.get("x-real-ip") ?? ""
    const ip = rawIp.includes(":") ? rawIp.split(":").slice(0, 4).join(":") : rawIp
    const rawZenApiKey = opts.parseApiKey(input.request.headers)
@@ -112,12 +104,10 @@ export async function handler(
    const projectId = input.request.headers.get("x-opencode-project") ?? ""
    const userAgent = input.request.headers.get("user-agent") ?? ""
    logger.metric({
-      is_stream: isStream,
      session: sessionId,
      request: requestId,
      client: ocClient,
      user_agent: userAgent,
-      "model.variant": variant,
      "model.tier": opts.modelList === "full" ? "zen" : "go",
    })
    const zenData = ZenData.list(opts.modelList)
@@ -175,7 +165,7 @@ export async function handler(
    )
    const providerBudget = await providerBudgetTracker?.check()

-    const retriableRequest = async (retry: RetryOptions = { excludeProviders: [], retryCount: 0 }) => {
+    const providerRequest = async () => {
      const providerInfo = selectProvider(
        model,
        zenData,
@@ -183,7 +173,6 @@ export async function handler(
        modelInfo,
        stickyId,
        trialProviders,
-        retry,
        stickyProvider,
        modelTpmLimits,
        modelTpsLimits,
@@ -199,80 +188,65 @@ export async function handler(
      })

      const startTimestamp = Date.now()
-      const reqUrl = providerInfo.modifyUrl(providerInfo.api, isStream)
-      const reqBody = JSON.stringify(
-        providerInfo.modifyBody({
-          ...createBodyConverter(opts.format, providerInfo.format)(body),
-          model: providerInfo.model,
-          ...(() => {
-            const replacer = (obj: Record<string, any>): Record<string, any> =>
-              Object.fromEntries(
-                Object.entries(obj).flatMap(([k, v]) => {
-                  if (Array.isArray(v)) return [[k, v]]
-                  if (typeof v === "object") return [[k, replacer(v)]]
-                  if (typeof v === "string") {
-                    if (v === "$workspace") return authInfo?.workspaceID ? [[k, authInfo.workspaceID]] : []
-                    if (v === "$org")
-                      return authInfo?.workspaceID ? [[k, authInfo.workspaceID.replace("wrk_", "org_")]] : []
-                    if (v === "$user") return stickyId ? [[k, stickyId]] : []
-                    if (v.startsWith("$header.")) {
-                      const headerValue = input.request.headers.get(v.slice(8))
-                      return headerValue ? [[k, headerValue]] : []
-                    }
-                  }
-                  return [[k, v]]
-                }),
-              )
-            return replacer(providerInfo.payloadModifier ?? {})
-          })(),
-        }),
-      )
+      const reqUrl = providerInfo.modifyUrl(providerInfo.api, googleStream ?? false)
+      const specialAnthropic =
+        providerInfo.format === "anthropic" &&
+        (providerInfo.model.startsWith("arn:aws:bedrock:") ||
+          providerInfo.model.startsWith("global.anthropic.") ||
+          providerInfo.model.startsWith("databricks-claude-"))
+      if (providerInfo.format !== opts.format) throw new Error("Zen provider format must match request format")
+      if (specialAnthropic) throw new Error("Anthropic provider body modifiers are incompatible with streaming")
+      const prepared = requestBody
+
+      const reqBody = (() => {
+        if (opts.format === "google") return body
+        if (!prepared) throw new Error("Missing prepared request body")
+        return prepared.stream(providerInfo.model, providerInfo.format === "oa-compat")
+      })()
      logger.debug("REQUEST URL: " + reqUrl)
-      logger.debug("REQUEST: " + reqBody.substring(0, 300) + "...")
+      logger.debug("REQUEST: " + (requestBody?.preview ?? "") + "...")
      const isNewInference =
        providerInfo.id.startsWith("console.") ||
        providerInfo.id.startsWith("console-go.") ||
        providerInfo.id.startsWith("inf.") ||
        providerInfo.id.startsWith("inf-go.")
-      const res = await fetchWithRetryableStatus(
-        reqUrl,
-        {
-          method: "POST",
-          headers: (() => {
-            const headers = new Headers(input.request.headers)
-            providerInfo.modifyHeaders(headers, providerInfo.apiKey, stickyId)
-            Object.entries(providerInfo.headerModifier ?? {}).forEach(([k, v]) => {
-              if (v === "$ip") return headers.set(k, ip)
-              if (v === "$caller") return headers.set(k, stickyId)
-              if (v === "$session") return headers.set(k, sessionId)
-              if (v === "$model") return headers.set(k, model)
-              if (v === "$request") return headers.set(k, requestId)
-              if (v === "$project") return headers.set(k, projectId)
-              if (v === "$workspace") {
-                if (authInfo?.workspaceID) headers.set(k, authInfo.workspaceID)
-                return
-              }
-              if (v === "$org") {
-                if (authInfo?.workspaceID) headers.set(k, authInfo.workspaceID.replace("wrk_", "org_"))
-                return
-              }
-              headers.set(k, v)
-            })
-            headers.delete("host")
-            headers.delete("content-length")
-            headers.delete("x-opencode-request")
-            if (!isNewInference) headers.delete("x-opencode-session")
-            headers.delete("x-opencode-project")
-            headers.delete("x-opencode-client")
-            return headers
-          })(),
-          body: reqBody,
-          // Propagate caller disconnects to the upstream provider request so
-          // abandoned Console requests do not leave orphaned inference work open.
-          signal: input.request.signal,
-        },
-        { count: isNewInference ? MAX_RETRYABLE_STATUS_RETRIES : 0 },
-      )
+      const res = await fetch(reqUrl, {
+        method: "POST",
+        headers: (() => {
+          const headers = new Headers(input.request.headers)
+          providerInfo.modifyHeaders(headers, providerInfo.apiKey, stickyId)
+          Object.entries(providerInfo.headerModifier ?? {}).forEach(([k, v]) => {
+            if (v === "$ip") return headers.set(k, ip)
+            if (v === "$caller") return headers.set(k, stickyId)
+            if (v === "$session") return headers.set(k, sessionId)
+            if (v === "$model") return headers.set(k, model)
+            if (v === "$request") return headers.set(k, requestId)
+            if (v === "$project") return headers.set(k, projectId)
+            if (v === "$workspace") {
+              if (authInfo?.workspaceID) headers.set(k, authInfo.workspaceID)
+              return
+            }
+            if (v === "$org") {
+              if (authInfo?.workspaceID) headers.set(k, authInfo.workspaceID.replace("wrk_", "org_"))
+              return
+            }
+            headers.set(k, v)
+          })
+          headers.delete("host")
+          headers.delete("content-length")
+          headers.delete("x-opencode-request")
+          if (!isNewInference) headers.delete("x-opencode-session")
+          headers.delete("x-opencode-project")
+          headers.delete("x-opencode-client")
+          return headers
+        })(),
+        body: reqBody,
+        // Propagate caller disconnects to the upstream provider request so
+        // abandoned Console requests do not leave orphaned inference work open.
+        signal: input.request.signal,
+      })
+      const isStream = res.headers.get("content-type")?.toLowerCase().includes("text/event-stream") ?? false
+      logger.metric({ is_stream: isStream })

      if (isNewInference) {
        const resEndpointId = res.headers.get("x-opencode-endpoint-id")
@@ -291,29 +265,10 @@ export async function handler(
        })
      }

-      // Try another provider => stop retrying if using fallback provider
-      if (
-        //!isNewInference &&
-        res.status !== 200 &&
-        // ie. 400 error is usually provider error like malformed request
-        res.status !== 400 &&
-        // ie. openai 404 error: Item with id 'msg_0ead8b004a3b165d0069436a6b6834819896da85b63b196a3f' not found.
-        !(modelInfo.id.startsWith("gpt-") && res.status === 404) &&
-        // ie. cannot change codex model providers mid-session
-        modelInfo.stickyProvider !== "strict" &&
-        modelInfo.fallbackProvider &&
-        providerInfo.id !== modelInfo.fallbackProvider
-      ) {
-        return retriableRequest({
-          excludeProviders: [...retry.excludeProviders, providerInfo.id],
-          retryCount: retry.retryCount + 1,
-        })
-      }
-
-      return { providerInfo, reqBody, res, startTimestamp }
+      return { providerInfo, res, startTimestamp, isStream }
    }

-    const { providerInfo, reqBody, res, startTimestamp } = await retriableRequest()
+    const { providerInfo, res, startTimestamp, isStream } = await providerRequest()

    // Store sticky provider
    if (res.status === 200) await stickyTracker?.set(providerInfo.id)
@@ -469,6 +424,8 @@ export async function handler(
      headers: resHeaders,
    })
  } catch (error: any) {
+    if (requestBody) void requestBody.cancel().catch(() => {})
+    else void input.request.body?.cancel().catch(() => {})
    // The caller disconnected before we finished. Because the outbound provider
    // request shares input.request.signal, an aborted caller surfaces here as an
    // AbortError. There is no client left to receive a body, so skip the error
@@ -593,7 +550,6 @@ export async function handler(
    modelInfo: ModelInfo,
    stickyId: string,
    trialProviders: string[] | undefined,
-    retry: RetryOptions,
    stickyProviderId: string | undefined,
    modelTpmLimits: Record<string, number> | undefined,
    modelTpsLimits: Record<string, { qualify: number; unqualify: number }> | undefined,
@@ -620,14 +576,11 @@ export async function handler(
        }))
      }

-      // Use fallback provider if max retries reached
      const fallbackProvider = allProviders.find((provider) => provider.id === modelInfo.fallbackProvider)
-      if (retry.retryCount === MAX_FAILOVER_RETRIES) return fallbackProvider

      let topPriority = Infinity
      const providers = allProviders
        .filter((provider) => provider.weight !== 0)
-        .filter((provider) => !retry.excludeProviders.includes(provider.id))
        .filter((provider) => {
          if (provider.budgetPriority === undefined) return true
          if (!providerBudget) return true
@@ -1035,15 +988,6 @@ export async function handler(
    providerInfo.apiKey = authInfo.provider.credentials
  }

-  async function fetchWithRetryableStatus(url: string, options: RequestInit, retry = { count: 0 }) {
-    const res = await fetch(url, options)
-    if ([429, 529].includes(res.status) && retry.count < MAX_RETRYABLE_STATUS_RETRIES) {
-      await new Promise((resolve) => setTimeout(resolve, Math.pow(2, retry.count) * 500))
-      return fetchWithRetryableStatus(url, options, { count: retry.count + 1 })
-    }
-    return res
-  }
-
  function calculateCost(modelInfo: ModelInfo, usageInfo: UsageInfo) {
    const { inputTokens, outputTokens, reasoningTokens, cacheReadTokens, cacheWrite5mTokens, cacheWrite1hTokens } =
      usageInfo
@@ -0,0 +1,191 @@
+const TAIL_LIMIT = 4 * 1024
+const encoder = new TextEncoder()
+
+export async function prepareRequestBody(body: ReadableStream<Uint8Array>) {
+  const reader = body.getReader()
+  const chunks: Uint8Array[] = []
+  const decoder = new TextDecoder()
+  let text = ""
+  let done = false
+  let searchFrom = 0
+  let bom = 0
+  let match: RegExpExecArray | null = null
+  const pattern = /("model"\s*:\s*")([^"]+)"/g
+
+  while (!done && !match) {
+    const next = await reader.read()
+    done = next.done
+    if (!next.value) continue
+    if (!chunks.length && next.value[0] === 0xef && next.value[1] === 0xbb && next.value[2] === 0xbf) bom = 3
+    chunks.push(next.value)
+    text += decoder.decode(next.value, { stream: true })
+    pattern.lastIndex = searchFrom
+    match = pattern.exec(text)
+    searchFrom = Math.max(0, text.length - 256)
+  }
+  if (done) {
+    text += decoder.decode()
+    if (!match) {
+      pattern.lastIndex = searchFrom
+      match = pattern.exec(text)
+    }
+  }
+
+  const found = (() => {
+    if (!match) return
+    const start = bom + utf8Length(text, match.index + match[1].length)
+    return { model: match[2], start, end: start + utf8Length(match[2], match[2].length) }
+  })()
+  const preview = text.substring(0, 300)
+  text = ""
+  match = null
+  let used = false
+
+  return {
+    model: found?.model ?? "",
+    preview,
+    cancel: () => reader.cancel(),
+    stream(providerModel: string, includeUsage: boolean) {
+      if (used) throw new Error("Request body stream already consumed")
+      if (!found) throw new Error("Missing model field")
+      used = true
+
+      const initial = replace(chunks, found.start, found.end, providerModel)
+      chunks.length = 0
+      const output = passthrough(initial, reader, done)
+      if (!includeUsage) return output
+      return appendUsage(output)
+    },
+  }
+}
+
+function utf8Length(value: string, end: number) {
+  let length = 0
+  for (let i = 0; i < end; i++) {
+    const code = value.charCodeAt(i)
+    if (code <= 0x7f) length++
+    else if (code <= 0x7ff) length += 2
+    else if (code >= 0xd800 && code <= 0xdbff && i + 1 < end && value.charCodeAt(i + 1) >= 0xdc00) {
+      length += 4
+      i++
+    } else length += 3
+  }
+  return length
+}
+
+function replace(chunks: Uint8Array[], start: number, end: number, value: string) {
+  let offset = 0
+  let inserted = false
+  return chunks.flatMap((chunk) => {
+    const chunkStart = offset
+    const chunkEnd = offset + chunk.length
+    offset = chunkEnd
+    if (chunkEnd <= start || chunkStart >= end) return [chunk]
+
+    const parts = [chunk.subarray(0, Math.max(0, start - chunkStart))]
+    if (!inserted) {
+      parts.push(encoder.encode(value))
+      inserted = true
+    }
+    parts.push(chunk.subarray(Math.min(chunk.length, end - chunkStart)))
+    return parts.filter((part) => part.length)
+  })
+}
+
+function passthrough(
+  initial: Array<Uint8Array | undefined>,
+  reader: ReadableStreamDefaultReader<Uint8Array>,
+  sourceDone: boolean,
+) {
+  let done = sourceDone
+  let index = 0
+  return new ReadableStream<Uint8Array>({
+    async pull(controller) {
+      const chunk = initial[index]
+      if (chunk) {
+        initial[index++] = undefined
+        controller.enqueue(chunk)
+        return
+      }
+      initial.length = 0
+      if (done) {
+        controller.close()
+        return
+      }
+      const next = await reader.read()
+      done = next.done
+      if (next.value) controller.enqueue(next.value)
+      if (done) controller.close()
+    },
+    cancel(reason) {
+      initial.length = 0
+      return reader.cancel(reason)
+    },
+  })
+}
+
+function appendUsage(body: ReadableStream<Uint8Array>) {
+  const reader = body.getReader()
+  const decoder = new TextDecoder()
+  let tail = new Uint8Array()
+  let streamText = ""
+  let isStream = false
+  const inspect = (chunk?: Uint8Array) => {
+    streamText += chunk ? decoder.decode(chunk, { stream: true }) : decoder.decode()
+    for (const match of streamText.matchAll(/"stream"\s*:\s*(true|false)/g)) isStream = match[1] === "true"
+    streamText = streamText.slice(-64)
+  }
+  return new ReadableStream<Uint8Array>({
+    async pull(controller) {
+      while (true) {
+        const next = await reader.read()
+        if (next.done) {
+          inspect()
+          if (!isStream) {
+            if (tail.length) controller.enqueue(tail)
+            controller.close()
+            return
+          }
+          const close = tail.lastIndexOf(125)
+          if (close < 0) {
+            controller.error(new Error("Invalid JSON request body"))
+            return
+          }
+          if (close) controller.enqueue(tail.subarray(0, close))
+          controller.enqueue(encoder.encode(',"stream_options":{"include_usage":true}}'))
+          if (close + 1 < tail.length) controller.enqueue(tail.subarray(close + 1))
+          controller.close()
+          return
+        }
+
+        const chunk = next.value
+        inspect(chunk)
+        if (tail.length + chunk.length <= TAIL_LIMIT) {
+          const combined = new Uint8Array(tail.length + chunk.length)
+          combined.set(tail)
+          combined.set(chunk, tail.length)
+          tail = combined
+          continue
+        }
+
+        const emit = tail.length + chunk.length - TAIL_LIMIT
+        if (emit <= tail.length) {
+          controller.enqueue(tail.subarray(0, emit))
+          const combined = new Uint8Array(TAIL_LIMIT)
+          combined.set(tail.subarray(emit))
+          combined.set(chunk, tail.length - emit)
+          tail = combined
+          return
+        }
+
+        if (tail.length) controller.enqueue(tail)
+        controller.enqueue(chunk.subarray(0, emit - tail.length))
+        tail = chunk.slice(emit - tail.length)
+        return
+      }
+    },
+    cancel(reason) {
+      return reader.cancel(reason)
+    },
+  })
+}
@@ -0,0 +1,122 @@
+import { describe, expect, test } from "bun:test"
+import { prepareRequestBody } from "../src/routes/zen/util/requestBody"
+
+describe("Zen request body streaming", () => {
+  test("patches the leading model without buffering the remaining body", async () => {
+    let reads = 0
+    const body = new ReadableStream<Uint8Array>(
+      {
+        pull(controller) {
+          const chunks = [
+            '{"model":"client-model","stream":true,"messages":[',
+            JSON.stringify({ role: "user", content: "large payload" }),
+            "]}",
+          ]
+          const chunk = chunks[reads++]
+          if (chunk) controller.enqueue(new TextEncoder().encode(chunk))
+          else controller.close()
+        },
+      },
+      { highWaterMark: 0 },
+    )
+
+    const request = await prepareRequestBody(body)
+    expect(request.model).toBe("client-model")
+    expect(reads).toBe(1)
+
+    const output = await new Response(request.stream("provider-model", false)).text()
+    expect(JSON.parse(output)).toEqual({
+      model: "provider-model",
+      stream: true,
+      messages: [{ role: "user", content: "large payload" }],
+    })
+  })
+
+  test("appends stream usage options at the end of the request", async () => {
+    const body = new Blob(['{"model":"client-model","stream":true,"messages":[]}   ']).stream()
+    const request = await prepareRequestBody(body)
+    const output = await new Response(request.stream("provider-model", true)).text()
+
+    expect(JSON.parse(output)).toEqual({
+      model: "provider-model",
+      stream: true,
+      messages: [],
+      stream_options: { include_usage: true },
+    })
+    expect(output.endsWith("   ")).toBe(true)
+  })
+
+  test("detects streaming after a large message while forwarding", async () => {
+    const content = "x".repeat(128 * 1024)
+    let reads = 0
+    const chunks = [
+      '{"model":"client-model","messages":[',
+      JSON.stringify({ role: "user", content }),
+      '],"stream":true}',
+    ]
+    const body = new ReadableStream<Uint8Array>(
+      {
+        pull(controller) {
+          const chunk = chunks[reads++]
+          if (chunk) controller.enqueue(new TextEncoder().encode(chunk))
+          else controller.close()
+        },
+      },
+      { highWaterMark: 0 },
+    )
+    const request = await prepareRequestBody(body)
+    expect(reads).toBe(1)
+    const output = await new Response(request.stream("provider-model", true)).text()
+
+    expect(JSON.parse(output)).toEqual({
+      model: "provider-model",
+      messages: [{ role: "user", content }],
+      stream: true,
+      stream_options: { include_usage: true },
+    })
+  })
+
+  test("buffers through a late model field and then streams the rest", async () => {
+    const content = "こんにちは".repeat(32 * 1024)
+    let reads = 0
+    const chunks = [
+      '{"messages":[',
+      JSON.stringify({ role: "user", content }),
+      '],"model":"client-model","stream":true,"extra":"after-model"}',
+    ]
+    const body = new ReadableStream<Uint8Array>(
+      {
+        pull(controller) {
+          const chunk = chunks[reads++]
+          if (chunk) controller.enqueue(new TextEncoder().encode(chunk))
+          else controller.close()
+        },
+      },
+      { highWaterMark: 0 },
+    )
+    const request = await prepareRequestBody(body)
+
+    expect(request.model).toBe("client-model")
+    expect(reads).toBe(3)
+    expect(JSON.parse(await new Response(request.stream("provider-model", true)).text())).toEqual({
+      messages: [{ role: "user", content }],
+      model: "provider-model",
+      stream: true,
+      extra: "after-model",
+      stream_options: { include_usage: true },
+    })
+  })
+
+  test("preserves a UTF-8 BOM while patching the model", async () => {
+    const body = new Blob(['\uFEFF{"messages":[],"model":"client-model","stream":false}']).stream()
+    const request = await prepareRequestBody(body)
+    const output = new Uint8Array(await new Response(request.stream("provider-model", false)).arrayBuffer())
+
+    expect([...output.subarray(0, 3)]).toEqual([0xef, 0xbb, 0xbf])
+    expect(JSON.parse(new TextDecoder().decode(output))).toEqual({
+      messages: [],
+      model: "provider-model",
+      stream: false,
+    })
+  })
+})
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/console-core",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -97,6 +97,18 @@ export namespace Workspace {
    },
  )

+  export const unblock = fn(
+    z.object({
+      workspaceID: Identifier.schema("workspace"),
+    }),
+    async (input) => {
+      const result = await Database.use((tx) =>
+        tx.update(WorkspaceTable).set({ is_blocked: false }).where(eq(WorkspaceTable.id, input.workspaceID)),
+      )
+      if (result.rowsAffected === 0) throw new Error("Workspace not found")
+    },
+  )
+
  export const remove = fn(z.void(), async () => {
    await Database.use((tx) =>
      tx
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-function",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "$schema": "https://json.schemastore.org/package.json",
  "private": true,
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-mail",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "dependencies": {
    "@jsx-email/all": "2.2.3",
    "@jsx-email/cli": "1.4.3",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-support",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "name": "@opencode-ai/core",
  "type": "module",
  "license": "MIT",
@@ -62,7 +62,7 @@
  },
  "dependencies": {
    "@ai-sdk/alibaba": "1.0.17",
-    "@ai-sdk/amazon-bedrock": "4.0.112",
+    "@ai-sdk/amazon-bedrock": "4.0.158",
    "@ai-sdk/anthropic": "3.0.82",
    "@ai-sdk/azure": "3.0.88",
    "@ai-sdk/cerebras": "2.0.41",
@@ -45,9 +45,18 @@ function oauth(http: HttpClient.HttpClient) {
    authorize: () =>
      Effect.gen(function* () {
        const device = yield* post(http, `${defaultServer}/auth/device/code`, { client_id: clientID }, Device)
+        const verification = yield* Effect.try({
+          try: () => {
+            const url = new URL(device.verification_uri_complete, `${defaultServer}/`)
+            if (url.protocol !== "http:" && url.protocol !== "https:") throw new Error("expected HTTP(S)")
+            return url
+          },
+          catch: (cause) =>
+            new Error(`Invalid device verification URL: ${cause instanceof Error ? cause.message : String(cause)}`),
+        })
        return {
          mode: "auto" as const,
-          url: `${defaultServer}${device.verification_uri_complete}`,
+          url: verification.href,
          instructions: `Enter code: ${device.user_code}`,
          callback: poll(http, defaultServer, device.device_code, Duration.seconds(device.interval)),
        }
@@ -1,5 +1,6 @@
 import { describe, expect } from "bun:test"
 import { Effect } from "effect"
+import { HttpClient, HttpClientResponse } from "effect/unstable/http"
 import { Catalog } from "@opencode-ai/core/catalog"
 import { Credential } from "@opencode-ai/core/credential"
 import { EventV2 } from "@opencode-ai/core/event"
@@ -14,14 +15,16 @@ import { PluginTestLayer } from "./fixture"

 const it = testEffect(PluginTestLayer)

-const addPlugin = Effect.fn(function* () {
+const addPlugin = Effect.fn(function* (http?: HttpClient.HttpClient) {
  const plugin = yield* PluginV2.Service
  const host = yield* PluginHost.make(plugin)
  const events = yield* EventV2.Service
  const integration = yield* Integration.Service
+  const client = yield* HttpClient.HttpClient
  yield* OpencodePlugin.effect(host).pipe(
    Effect.provideService(EventV2.Service, events),
    Effect.provideService(Integration.Service, integration),
+    Effect.provideService(HttpClient.HttpClient, http ?? client),
  )
 })

@@ -82,6 +85,63 @@ describe("OpencodePlugin", () => {
    }),
  )

+  it.effect("resolves origin-rooted device verification URLs", () =>
+    Effect.gen(function* () {
+      const http = HttpClient.make((request) =>
+        Effect.succeed(
+          HttpClientResponse.fromWeb(
+            request,
+            Response.json({
+              device_code: "device",
+              user_code: "user",
+              verification_uri_complete: "/console/device?user_code=user&client_id=opencode-cli",
+              expires_in: 60,
+              interval: 60,
+            }),
+          ),
+        ),
+      )
+      yield* addPlugin(http)
+      const integration = yield* Integration.Service
+      const attempt = yield* integration.connection.oauth({
+        integrationID: Integration.ID.make("opencode"),
+        methodID: Integration.MethodID.make("device"),
+        inputs: {},
+      })
+      expect(attempt.url).toBe("https://opencode.ai/console/device?user_code=user&client_id=opencode-cli")
+    }),
+  )
+
+  it.effect("rejects malformed device verification URLs", () =>
+    Effect.gen(function* () {
+      const http = HttpClient.make((request) =>
+        Effect.succeed(
+          HttpClientResponse.fromWeb(
+            request,
+            Response.json({
+              device_code: "device",
+              user_code: "user",
+              verification_uri_complete: "http://[::1",
+              expires_in: 60,
+              interval: 60,
+            }),
+          ),
+        ),
+      )
+      yield* addPlugin(http)
+      const integration = yield* Integration.Service
+      const error = yield* integration.connection
+        .oauth({
+          integrationID: Integration.ID.make("opencode"),
+          methodID: Integration.MethodID.make("device"),
+          inputs: {},
+        })
+        .pipe(Effect.flip)
+      expect(error).toBeInstanceOf(Integration.AuthorizationError)
+      expect(String(error.cause)).toContain("Invalid device verification URL")
+    }),
+  )
+
  it.live("loads providers and models from the connected OpenCode server", () =>
    Effect.acquireUseRelease(
      Effect.sync(() => {
@@ -1,7 +1,7 @@
 {
  "name": "@opencode-ai/desktop",
  "private": true,
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "homepage": "https://opencode.ai",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "name": "@opencode-ai/effect-drizzle-sqlite",
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "name": "@opencode-ai/effect-sqlite-node",
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/enterprise",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/function",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "$schema": "https://json.schemastore.org/package.json",
  "private": true,
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "name": "@opencode-ai/http-recorder",
  "description": "Record and replay Effect HTTP client traffic with deterministic cassettes",
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "name": "@opencode-ai/llm",
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "name": "opencode",
  "type": "module",
  "license": "MIT",
@@ -56,7 +56,7 @@
    "@actions/github": "6.0.1",
    "@agentclientprotocol/sdk": "0.21.0",
    "@ai-sdk/alibaba": "1.0.17",
-    "@ai-sdk/amazon-bedrock": "4.0.112",
+    "@ai-sdk/amazon-bedrock": "4.0.158",
    "@ai-sdk/anthropic": "3.0.82",
    "@ai-sdk/azure": "3.0.88",
    "@ai-sdk/cerebras": "2.0.60",
@@ -396,10 +396,18 @@ const layer: Layer.Layer<Service, never, AccountRepo.Service | HttpClient.HttpCl
      const parsed = yield* HttpClientResponse.schemaBodyJson(DeviceAuth)(response).pipe(
        mapAccountServiceError("Failed to decode response"),
      )
+      const verification = yield* Effect.try({
+        try: () => {
+          const url = new URL(parsed.verification_uri_complete, `${normalizedServer}/`)
+          if (url.protocol !== "http:" && url.protocol !== "https:") throw new Error("expected HTTP(S)")
+          return url.href
+        },
+        catch: (cause) => new AccountServiceError({ message: "Invalid device verification URL", cause }),
+      })
      return new Login({
        code: parsed.device_code,
        user: parsed.user_code,
-        url: `${normalizedServer}${parsed.verification_uri_complete}`,
+        url: verification,
        server: normalizedServer,
        expiry: parsed.expires_in,
        interval: parsed.interval,
@@ -98,6 +98,12 @@ function googleVertexAnthropicBaseURL(project: string | undefined, location: str
  return `https://aiplatform.${location}.rep.googleapis.com/v1/projects/${project}/locations/${location}/publishers/anthropic/models`
 }

+function googleVertexEndpoint(location: string) {
+  if (location === "global") return "aiplatform.googleapis.com"
+  if (location === "eu" || location === "us") return `aiplatform.${location}.rep.googleapis.com`
+  return `${location}-aiplatform.googleapis.com`
+}
+
 type BundledSDK = {
  languageModel(modelId: string): LanguageModelV3
  chat?: (modelId: string) => LanguageModelV3
@@ -519,11 +525,10 @@ function custom(dep: CustomDep): Record<string, CustomLoader> {
      return {
        autoload: true,
        vars(_options: Record<string, any>) {
-          const endpoint = location === "global" ? "aiplatform.googleapis.com" : `${location}-aiplatform.googleapis.com`
          return {
            ...(project && { GOOGLE_VERTEX_PROJECT: project }),
            GOOGLE_VERTEX_LOCATION: location,
-            GOOGLE_VERTEX_ENDPOINT: endpoint,
+            GOOGLE_VERTEX_ENDPOINT: googleVertexEndpoint(location),
          }
        },
        options: {
@@ -1303,13 +1303,13 @@ export function options(input: {
      }
    }

-    // Only set textVerbosity for non-chat gpt-5.x models
-    // Chat models (e.g. gpt-5.2-chat-latest) only support "medium" verbosity
+    // Generic OpenAI-compatible APIs do not necessarily support OpenAI's verbosity parameter.
+    // Only enable the default for integrations known to implement it.
    if (
      input.model.api.id.includes("gpt-5.") &&
      !input.model.api.id.includes("codex") &&
      !input.model.api.id.includes("-chat") &&
-      input.model.providerID !== "azure"
+      (input.model.api.npm === "@ai-sdk/openai" || input.model.api.npm === "@ai-sdk/amazon-bedrock/mantle")
    ) {
      result["textVerbosity"] = "low"
    }
@@ -29,8 +29,6 @@ import * as OtelTracer from "@effect/opentelemetry/Tracer"
 import { LLMAISDK } from "./llm/ai-sdk"
 import { LLMNativeRuntime } from "./llm/native-runtime"
 import { LLMRequestPrep } from "./llm/request"
-import { ProviderError } from "@/provider/error"
-import type { LanguageModelV3StreamPart } from "@ai-sdk/provider"

 export const OUTPUT_TOKEN_MAX = ProviderTransform.OUTPUT_TOKEN_MAX

@@ -340,26 +338,6 @@ const live: Layer.Layer<
                  }
                  return args.params
                },
-                async wrapStream({ doStream }) {
-                  const result = await doStream()
-                  return {
-                    ...result,
-                    stream: result.stream.pipeThrough(
-                      new TransformStream<LanguageModelV3StreamPart, LanguageModelV3StreamPart>({
-                        transform(part, controller) {
-                          if (part.type === "finish" && part.finishReason.raw === "network_error") {
-                            controller.enqueue({
-                              type: "error",
-                              error: new ProviderError.ResponseStreamError("Provider finish_reason: network_error"),
-                            })
-                            return
-                          }
-                          controller.enqueue(part)
-                        },
-                      }),
-                    ),
-                  }
-                },
              },
            ],
          }),
@@ -2,6 +2,7 @@ import { FinishReason, LLMEvent, ProviderMetadata, ToolResultValue } from "@open
 import { Effect, Schema } from "effect"
 import { type streamText } from "ai"
 import { errorMessage } from "@/util/error"
+import { ProviderError } from "@/provider/error"

 type Result = Awaited<ReturnType<typeof streamText>>
 type AISDKEvent = Result["fullStream"] extends AsyncIterable<infer T> ? T : never
@@ -85,6 +86,8 @@ export function toLLMEvents(
      return Effect.succeed([LLMEvent.stepStart({ index: state.step })])

    case "finish-step":
+      if (event.rawFinishReason === "network_error")
+        return Effect.fail(new ProviderError.ResponseStreamError("Provider finish_reason: network_error"))
      return Effect.sync(() => {
        const original = providerMetadata(event.providerMetadata)
        const metadata =
@@ -1110,7 +1110,7 @@ const layer = Layer.effect(

          if (
            lastAssistant?.finish &&
-            !["tool-calls"].includes(lastAssistant.finish) &&
+            !["tool-calls", "unknown"].includes(lastAssistant.finish) &&
            !hasToolCalls &&
            lastAssistant.parentID === lastUser.id
          ) {
@@ -217,6 +217,10 @@ export const TaskTool = Tool.define(
              : result.info.error.name
          return yield* Effect.fail(new Error(`Subagent failed (task_id: ${nextSession.id}): ${message}`))
        }
+        const failed = result.parts.findLast((item) => item.type === "tool" && item.state.status === "error")
+        if (failed?.type === "tool" && failed.state.status === "error") {
+          return yield* Effect.fail(new Error(`Subagent failed (task_id: ${nextSession.id}): ${failed.state.error}`))
+        }
        return result.parts.findLast((item) => item.type === "text")?.text ?? ""
      })

@@ -10,6 +10,7 @@ import { Account } from "../../src/account/account"
 import {
  AccessToken,
  AccountID,
+  AccountServiceError,
  AccountTransportError,
  DeviceCode,
  Login,
@@ -71,18 +72,18 @@ const deviceTokenClient = (body: unknown, status = 400) =>
 const poll = (body: unknown, status = 400) =>
  Account.Service.use((s) => s.poll(login())).pipe(Effect.provide(live(deviceTokenClient(body, status))))

-it.live("login normalizes trailing slashes in the provided server URL", () =>
+it.live("login resolves origin-rooted verification URLs from servers with base paths", () =>
  Effect.gen(function* () {
    const seen: Array<string> = []
    const client = HttpClient.make((req) =>
      Effect.gen(function* () {
        seen.push(`${req.method} ${req.url}`)

-        if (req.url === "https://one.example.com/auth/device/code") {
+        if (req.url === "https://one.example.com/console/auth/device/code") {
          return json(req, {
            device_code: "device-code",
            user_code: "user-code",
-            verification_uri_complete: "/device?user_code=user-code",
+            verification_uri_complete: "/console/device?user_code=user-code",
            expires_in: 600,
            interval: 5,
          })
@@ -92,11 +93,31 @@ it.live("login normalizes trailing slashes in the provided server URL", () =>
      }),
    )

-    const result = yield* Account.use.login("https://one.example.com/").pipe(Effect.provide(live(client)))
+    const result = yield* Account.use.login("https://one.example.com/console/").pipe(Effect.provide(live(client)))

-    expect(seen).toEqual(["POST https://one.example.com/auth/device/code"])
-    expect(result.server).toBe("https://one.example.com")
-    expect(result.url).toBe("https://one.example.com/device?user_code=user-code")
+    expect(seen).toEqual(["POST https://one.example.com/console/auth/device/code"])
+    expect(result.server).toBe("https://one.example.com/console")
+    expect(result.url).toBe("https://one.example.com/console/device?user_code=user-code")
+  }),
+)
+
+it.live("login rejects malformed device verification URLs", () =>
+  Effect.gen(function* () {
+    const client = HttpClient.make((req) =>
+      Effect.succeed(
+        json(req, {
+          device_code: "device-code",
+          user_code: "user-code",
+          verification_uri_complete: "http://[::1",
+          expires_in: 600,
+          interval: 5,
+        }),
+      ),
+    )
+
+    const error = yield* Effect.flip(Account.use.login("https://one.example.com").pipe(Effect.provide(live(client))))
+    expect(error).toBeInstanceOf(AccountServiceError)
+    if (error instanceof AccountServiceError) expect(error.message).toBe("Invalid device verification URL")
  }),
 )

@@ -82,10 +82,10 @@ describe("opencode run (non-interactive subprocess)", () => {
  )

  // The test provider's SSE error item is interpreted by the SDK as an unknown
-  // finish, not a fatal provider/session error. Lock that distinction in so it
-  // is not accidentally used as the failure compatibility oracle.
+  // finish, not a fatal provider/session error. Unknown finishes should continue
+  // the prompt loop so a subsequent response can complete the run.
  cliIt.concurrent(
-    "unknown stream finish preserves partial output and exits 0",
+    "unknown stream finish preserves partial output and continues",
    ({ llm, opencode }) =>
      Effect.gen(function* () {
        yield* llm.push(
@@ -95,9 +95,10 @@ describe("opencode run (non-interactive subprocess)", () => {
          }),
        )
        yield* llm.fail("upstream provider exploded mid-stream")
+        yield* llm.text("recovered")
        const result = yield* opencode.run("trigger midstream error", { timeoutMs: 30_000 })
        expect(result.exitCode).toBe(0)
-        expect(result.stdout).toBe("partial response\n")
+        expect(result.stdout).toBe("partial response\nrecovered\n")
        expect(result.stderr).not.toContain("upstream provider exploded mid-stream")
      }),
    60_000,
@@ -213,7 +214,7 @@ describe("opencode run (non-interactive subprocess)", () => {
  )

  cliIt.concurrent(
-    "--format json records partial output for an unknown stream finish",
+    "--format json records an unknown stream finish and continuation",
    ({ llm, opencode }) =>
      Effect.gen(function* () {
        yield* llm.push(
@@ -223,6 +224,7 @@ describe("opencode run (non-interactive subprocess)", () => {
          }),
        )
        yield* llm.fail("provider failed")
+        yield* llm.text("recovered")
        const result = yield* opencode.run("fail after output", { format: "json" })

        const events = opencode.parseJsonEvents(result.stdout)
@@ -234,9 +236,14 @@ describe("opencode run (non-interactive subprocess)", () => {
          "step_finish",
          "step_start",
          "step_finish",
+          "step_start",
+          "text",
+          "step_finish",
        ])
        expect(events[1]?.part).toEqual(expect.objectContaining({ type: "text", text: "partial json" }))
-        expect(events.at(-1)?.part).toEqual(expect.objectContaining({ type: "step-finish", reason: "unknown" }))
+        expect(events[5]?.part).toEqual(expect.objectContaining({ type: "step-finish", reason: "unknown" }))
+        expect(events[7]?.part).toEqual(expect.objectContaining({ type: "text", text: "recovered" }))
+        expect(events.at(-1)?.part).toEqual(expect.objectContaining({ type: "step-finish", reason: "stop" }))
      }),
    60_000,
  )
@@ -277,50 +284,6 @@ describe("opencode run (non-interactive subprocess)", () => {
    60_000,
  )

-  cliIt.concurrent(
-    "answers requested permissions from subagents",
-    ({ llm, opencode }) =>
-      Effect.gen(function* () {
-        yield* llm.tool("task", {
-          description: "Run a command",
-          prompt: "Run the requested command",
-          subagent_type: "general",
-        })
-        yield* llm.tool("bash", { command: "printf child", description: "Print from the child" })
-        yield* llm.text("child finished")
-        yield* llm.text("parent finished")
-
-        const result = yield* opencode.run("delegate a command", {
-          permission: { bash: "ask" },
-          extraArgs: ["--dangerously-skip-permissions"],
-          timeoutMs: 10_000,
-        })
-
-        opencode.expectExit(result, 0)
-        expect(result.stdout).toContain("parent finished")
-
-        yield* llm.reset
-        yield* llm.tool("task", {
-          description: "Run a command",
-          prompt: "Run the requested command",
-          subagent_type: "general",
-        })
-        yield* llm.tool("bash", { command: "printf child", description: "Print from the child" })
-        yield* llm.text("child continued")
-        yield* llm.text("parent continued")
-
-        const rejected = yield* opencode.run("delegate a rejected command", {
-          permission: { bash: "ask" },
-          timeoutMs: 10_000,
-        })
-
-        opencode.expectExit(rejected, 0)
-        expect(rejected.stderr).toContain("permission requested: bash")
-        expect(rejected.stdout).toContain("child continued")
-      }),
-    30_000,
-  )
-
  cliIt.live(
    "attach mode sends client-local file contents without a shared path",
    ({ home, llm, opencode }) =>
@@ -1909,6 +1909,32 @@ it.instance("Google Vertex: keeps regional Claude endpoints unchanged", () =>
  }),
 )

+it.instance("Google Vertex: uses REP endpoint for Gemini continental multi-regions", () =>
+  Effect.gen(function* () {
+    yield* set("GOOGLE_CLOUD_PROJECT", "test-project")
+    yield* set("VERTEX_LOCATION", "eu")
+    const provider = yield* Provider.Service
+    const model = yield* provider.getModel(ProviderV2.ID.make("google-vertex"), ModelV2.ID.make("gemini-3.5-flash"))
+    const language = yield* provider.getLanguage(model)
+    expect(languageBaseURL(language)).toBe(
+      "https://aiplatform.eu.rep.googleapis.com/v1beta1/projects/test-project/locations/eu/publishers/google",
+    )
+  }),
+)
+
+it.instance("Google Vertex: keeps regional Gemini endpoints unchanged", () =>
+  Effect.gen(function* () {
+    yield* set("GOOGLE_CLOUD_PROJECT", "test-project")
+    yield* set("VERTEX_LOCATION", "europe-west1")
+    const provider = yield* Provider.Service
+    const model = yield* provider.getModel(ProviderV2.ID.make("google-vertex"), ModelV2.ID.make("gemini-3.5-flash"))
+    const language = yield* provider.getLanguage(model)
+    expect(languageBaseURL(language)).toBe(
+      "https://europe-west1-aiplatform.googleapis.com/v1beta1/projects/test-project/locations/europe-west1/publishers/google",
+    )
+  }),
+)
+
 it.instance("cloudflare-ai-gateway loads with env variables", () =>
  Effect.gen(function* () {
    yield* set("CLOUDFLARE_ACCOUNT_ID", "test-account")
@@ -523,6 +523,7 @@ describe("ProviderTransform.options - gpt-5 textVerbosity", () => {
    expect(result.reasoningEffort).toBe("medium")
    expect(result.reasoningSummary).toBeUndefined()
    expect(result.include).toBeUndefined()
+    expect(result.textVerbosity).toBeUndefined()
  })

  test("azure chat completions omit Responses-only reasoning options after variants merge", async () => {
@@ -651,6 +651,63 @@ it.live("session.processor effect tests retry OpenAI-compatible midstream server
  ),
 )

+it.live("session.processor effect tests retry network_error finish reasons", () =>
+  provideTmpdirServer(
+    ({ dir, llm }) =>
+      Effect.gen(function* () {
+        const { processors, session, provider } = yield* boot()
+
+        yield* llm.push(
+          raw({
+            chunks: [
+              {
+                id: "chatcmpl-network-error",
+                object: "chat.completion.chunk",
+                choices: [{ index: 0, delta: { role: "assistant", content: "" }, finish_reason: "network_error" }],
+              },
+            ],
+          }),
+        )
+        yield* llm.text("after retry")
+
+        const chat = yield* session.create({})
+        const parent = yield* user(chat.id, "retry network error")
+        const msg = yield* assistant(chat.id, parent.id, path.resolve(dir))
+        const mdl = yield* provider.getModel(ref.providerID, ref.modelID)
+        const handle = yield* processors.create({
+          assistantMessage: msg,
+          sessionID: chat.id,
+          model: mdl,
+        })
+
+        const value = yield* handle.process({
+          user: {
+            id: parent.id,
+            sessionID: chat.id,
+            role: "user",
+            time: parent.time,
+            agent: parent.agent,
+            model: { providerID: ref.providerID, modelID: ref.modelID },
+          } satisfies SessionV1.User,
+          sessionID: chat.id,
+          model: mdl,
+          agent: agent(),
+          system: [],
+          messages: [{ role: "user", content: "retry network error" }],
+          tools: {},
+        })
+
+        const parts = yield* MessageV2.parts(msg.id)
+
+        expect(value).toBe("continue")
+        expect(yield* llm.calls).toBe(2)
+        expect(parts.some((part) => part.type === "text" && part.text === "after retry")).toBe(true)
+        expect(handle.message.error).toBeUndefined()
+      }),
+    { config: (url) => providerCfg(url) },
+  ),
+)
+
 it.live("session.processor effect tests publish retry status updates", () =>
  provideTmpdirServer(
    ({ dir, llm }) =>
@@ -850,6 +850,34 @@ it.instance("loop continues when finish is tool-calls", () =>
  }),
 )

+it.instance("loop continues when finish is unknown", () =>
+  Effect.gen(function* () {
+    const { llm } = yield* useServerConfig(providerCfg)
+    const prompt = yield* SessionPrompt.Service
+    const sessions = yield* Session.Service
+    const session = yield* sessions.create({
+      title: "Pinned",
+      permission: [{ permission: "*", pattern: "*", action: "allow" }],
+    })
+    yield* prompt.prompt({
+      sessionID: session.id,
+      agent: "build",
+      noReply: true,
+      parts: [{ type: "text", text: "hello" }],
+    })
+    yield* llm.push(reply())
+    yield* llm.text("second")
+
+    const result = yield* prompt.loop({ sessionID: session.id })
+    expect(yield* llm.calls).toBe(2)
+    expect(result.info.role).toBe("assistant")
+    if (result.info.role === "assistant") {
+      expect(result.parts.some((part) => part.type === "text" && part.text === "second")).toBe(true)
+      expect(result.info.finish).toBe("stop")
+    }
+  }),
+)
+
 it.instance("glob tool keeps instance context during prompt runs", () =>
  Effect.gen(function* () {
    const { dir, llm } = yield* useServerConfig(providerCfg)
@@ -100,6 +100,7 @@ function stubOps(opts?: {
  onPrompt?: (input: SessionPrompt.PromptInput) => void
  text?: string
  error?: NonNullable<SessionV1.Assistant["error"]>
+  toolError?: string
 }): TaskPromptOps {
  return {
    cancel: () => Effect.void,
@@ -107,7 +108,7 @@ function stubOps(opts?: {
    prompt: (input) =>
      Effect.sync(() => {
        opts?.onPrompt?.(input)
-        return reply(input, opts?.text ?? "done", opts?.error)
+        return reply(input, opts?.text ?? "done", opts?.error, opts?.toolError)
      }),
  }
 }
@@ -116,6 +117,7 @@ function reply(
  input: SessionPrompt.PromptInput,
  text: string,
  error?: NonNullable<SessionV1.Assistant["error"]>,
+  toolError?: string,
 ): SessionV1.WithParts {
  const id = MessageID.ascending()
  return {
@@ -143,6 +145,24 @@ function reply(
        type: "text",
        text,
      },
+      ...(toolError
+        ? [
+            {
+              id: PartID.ascending(),
+              messageID: id,
+              sessionID: input.sessionID,
+              type: "tool" as const,
+              tool: "read",
+              callID: "call-1",
+              state: {
+                status: "error" as const,
+                input: { filePath: "/external" },
+                error: toolError,
+                time: { start: Date.now(), end: Date.now() },
+              },
+            },
+          ]
+        : []),
    ],
  }
 }
@@ -307,6 +327,50 @@ describe("tool.task", () => {
    }),
  )

+  it.instance("execute surfaces terminal child tool errors with a resumable task_id", () =>
+    Effect.gen(function* () {
+      const sessions = yield* Session.Service
+      const { chat, assistant } = yield* seed()
+      const tool = yield* TaskTool
+      const def = yield* tool.init()
+
+      const exit = yield* def
+        .execute(
+          {
+            description: "inspect external directory",
+            prompt: "read the external directory",
+            subagent_type: "general",
+          },
+          {
+            sessionID: chat.id,
+            messageID: assistant.id,
+            agent: "build",
+            abort: new AbortController().signal,
+            extra: {
+              promptOps: stubOps({
+                text: "I will inspect the directory.",
+                toolError: "The user rejected permission to use this specific tool call.",
+              }),
+            },
+            messages: [],
+            metadata: () => Effect.void,
+            ask: () => Effect.void,
+          },
+        )
+        .pipe(Effect.exit)
+
+      expect(Exit.isFailure(exit)).toBe(true)
+      if (Exit.isSuccess(exit)) throw new Error("expected task failure")
+      const child = (yield* sessions.children(chat.id))[0]
+      const failure = Cause.squash(exit.cause)
+      expect(failure).toBeInstanceOf(Error)
+      if (!(failure instanceof Error)) throw new Error("expected Error defect")
+      expect(failure.message).toBe(
+        `Subagent failed (task_id: ${child?.id}): The user rejected permission to use this specific tool call.`,
+      )
+    }),
+  )
+
  it.instance("execute asks by default and skips checks when bypassed", () =>
    Effect.gen(function* () {
      const { chat, assistant } = yield* seed()
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/plugin",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/sdk",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/server",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/session-ui",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/slack",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/stats-app",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -812,7 +812,8 @@ function ModelTrendSection(props: {
                  <div data-slot="tooltip-divider" />
                  <p>
                    <span data-slot="tooltip-label">
-                      <i data-kind={props.lineTone === "active" ? "users" : "tokens"} /> {props.rowLabel}
+                      <i data-kind={props.lineTone === "active" ? "users" : "tokens"} />
+                      <span data-slot="tooltip-name">{props.rowLabel}</span>
                    </span>
                    <b>{props.formatValue(props.value(active.point))}</b>
                  </p>
@@ -473,13 +473,15 @@ function LabUsageSection(props: { lab: ModelCatalogLab; data: StatsLabData | nul
                    <div data-slot="tooltip-divider" />
                    <p>
                      <span data-slot="tooltip-label">
-                        <i data-kind="tokens" /> {i18n.t("lab.dailyTokens")}
+                        <i data-kind="tokens" />
+                        <span data-slot="tooltip-name">{i18n.t("lab.dailyTokens")}</span>
                      </span>
                      <b>{formatTokens(active.point.tokens)}</b>
                    </p>
                    <p>
                      <span data-slot="tooltip-label">
-                        <i data-kind="users" /> {i18n.t("model.uniqueUsers")}
+                        <i data-kind="users" />
+                        <span data-slot="tooltip-name">{i18n.t("model.uniqueUsers")}</span>
                      </span>
                      <b>{formatUsers(active.point.users)}</b>
                    </p>
@@ -1590,6 +1590,13 @@ body {
  white-space: nowrap;
 }

+[data-page="stats"] [data-component="chart-tooltip"] [data-slot="tooltip-name"] {
+  min-width: 0;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
 [data-page="stats"] [data-component="chart-tooltip"] [data-slot="tooltip-divider"] {
  height: 1px;
  margin: 4px -16px 2px;
@@ -1677,7 +1684,7 @@ body {

 [data-page="stats"] :is([data-section="top-models"], [data-section="unique-users"]) [data-component="chart-tooltip"] p {
  grid-template-columns: minmax(0, 1fr) auto;
-  gap: 4px;
+  gap: 1ch;
  height: 16px;
  margin: 4px 0 0;
  padding: 0 8px;
@@ -4723,7 +4730,7 @@ body {

 [data-page="stats"] [data-component="model-usage-chart"] [data-component="chart-tooltip"] p {
  grid-template-columns: minmax(0, 1fr) auto;
-  gap: 4px;
+  gap: 1ch;
  height: 20px;
  margin: 8px 0;
  padding: 0 8px;
@@ -568,8 +568,8 @@ function TopModelsChart(props: {
                              style={{
                                background: getRankColor(item.segment.model, item.index, segmentOrder(), usageColors),
                              }}
-                            />{" "}
-                            {item.segment.model}
+                            />
+                            <span data-slot="tooltip-name">{item.segment.model}</span>
                          </span>
                          <b>{formatUsageChartValue(item.segment.value, metric())}</b>
                        </p>
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/stats-core",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -4,6 +4,7 @@ import { modelAuthor, normalizeInferenceModel, statModel, statProvider } from ".

 describe("inference stat normalization", () => {
  test("normalizes model suffixes used by router/provider variants", () => {
+    expect(normalizeInferenceModel("GPT-5-Free")).toBe("gpt-5")
    expect(normalizeInferenceModel("deepseek-v4-flash-free")).toBe("deepseek-v4-flash")
    expect(normalizeInferenceModel("deepseek-v4-flash:global")).toBe("deepseek-v4-flash")
    expect(normalizeInferenceModel("mimo-v2.5-free")).toBe("mimo-v2.5")
@@ -41,6 +42,17 @@ describe("inference stat normalization", () => {
    expect(statProvider("unknown", "", "custom-provider")).toBe("custom-provider")
  })

+  test("merges renamed models under their current name", () => {
+    expect(statModel("x-preview-f", "")).toBe("ox-alpha")
+    expect(statModel("xiaomi/mimo-v2.5", "")).toBe("mimo-v2.5")
+    expect(toModelAggregate(aggregate("x-preview-f", "openai"))).toMatchObject([
+      {
+        provider: "openai",
+        model: "ox-alpha",
+      },
+    ])
+  })
+
  test("model aggregates prefer provider.model and use normalized model", () => {
    expect(toModelAggregate(aggregate("alpha-gpt-next", "openai"))).toEqual([])

@@ -6,6 +6,7 @@ import {
  EXCLUDED_MODELS,
  FREE_MODELS,
  MODEL_AUTHOR_RULES,
+  MODEL_NAME_ALIASES,
  RETIRED_STAT_PROVIDERS,
  statModel,
  statProvider,
@@ -302,6 +303,9 @@ function statPeriods(grain: "day" | "week", periodStart: Date, periodEnd: Date)
 function statModelSql(model: string, providerModel: string) {
  return `COALESCE(NULLIF(regexp_replace(CASE
      WHEN lower(${model}) = 'big-pickle' THEN regexp_replace(NULLIF(${providerModel}, ''), '^.*/', '')
+${Object.entries(MODEL_NAME_ALIASES)
+  .map(([from, to]) => `      WHEN lower(${model}) = ${sqlString(from)} THEN ${sqlString(to)}`)
+  .join("\n")}
      ELSE ${model}
    END, '(-free|:free|:global)+$', ''), ''), 'unknown')`
 }
@@ -15,11 +15,15 @@ export const MODEL_AUTHOR_RULES = [
 ] as const
 export const EXCLUDED_MODELS = new Set(["alpha-gpt-next"])
 export const FREE_MODELS = new Set(["gpt-5-nano", "grok-code", "big-pickle"])
-export const RETIRED_STAT_MODELS = ["big-pickle"]
+export const MODEL_NAME_ALIASES: Record<string, string> = {
+  "x-preview-f": "ox-alpha",
+  "xiaomi/mimo-v2.5": "mimo-v2.5",
+}
+export const RETIRED_STAT_MODELS = ["big-pickle", ...Object.keys(MODEL_NAME_ALIASES)]
 export const RETIRED_STAT_PROVIDERS = ["opencode"]

 export function normalizeInferenceModel(value: string | undefined) {
-  return (value || "unknown").replace(/(-free|:free|:global)+$/, "") || "unknown"
+  return (value || "unknown").toLowerCase().replace(/(-free|:free|:global)+$/, "") || "unknown"
 }

 export function modelAuthor(value: string | undefined) {
@@ -31,6 +35,8 @@ export function modelAuthor(value: string | undefined) {

 export function statModel(model: string | undefined, providerModel: string | undefined) {
  const normalized = normalizeInferenceModel(model)
+  const alias = MODEL_NAME_ALIASES[normalized.toLowerCase()]
+  if (alias) return alias
  if (RETIRED_STAT_MODELS.includes(normalized.toLowerCase()))
    return normalizeInferenceModel(providerModel?.split("/").at(-1))
  return normalized
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/stats-server",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/tui",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/ui",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "type": "module",
  "license": "MIT",
  "repository": {
@@ -2,7 +2,7 @@
  "name": "@opencode-ai/web",
  "type": "module",
  "license": "MIT",
-  "version": "1.18.19",
+  "version": "1.18.21",
  "scripts": {
    "dev": "astro dev",
    "dev:remote": "VITE_API_URL=https://api.opencode.ai astro dev",
@@ -68,6 +68,7 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال
 - **Qwen3.6 Plus**
 - **DeepSeek V4 Pro**
 - **DeepSeek V4 Flash**
+- **DeepSeek V4 Flash Vision Exp**
 - **Hy3**
 - **Ox Alpha Free** (لفترة محدودة)

@@ -87,29 +88,30 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال

 يوضح الجدول أدناه عددًا تقديريًا للطلبات بناءً على أنماط استخدام Go المعتادة:

-| Model                      | الطلبات لكل 5 ساعات | الطلبات في الأسبوع | الطلبات في الشهر |
-| -------------------------- | ------------------- | ------------------ | ---------------- |
-| Grok 4.5                   | 120                 | 300                | 600              |
-| GPT 5.6 Luna               | 2,050               | 5,100              | 10,250           |
-| GLM-5.3                    | 220                 | 540                | 1,080            |
-| GLM-5.2                    | 880                 | 2,150              | 4,300            |
-| GLM-5.1                    | 880                 | 2,150              | 4,300            |
-| Kimi K3                    | 110                 | 250                | 490              |
-| Kimi K2.7 Code             | 1,350               | 3,380              | 6,750            |
-| Kimi K2.6                  | 1,150               | 2,880              | 5,750            |
-| MiMo-V2.5                  | 30,100              | 75,200             | 150,400          |
-| MiMo-V2.5-Pro              | 3,250               | 8,150              | 16,300           |
-| MiniMax M3                 | 3,200               | 8,000              | 16,000           |
-| MiniMax M2.7               | 3,400               | 8,500              | 17,000           |
-| Muse Spark 1.2 Contributor | 45,300              | 113,300            | 226,600          |
-| Qwen3.8 Max                | 160                 | 400                | 810              |
-| Qwen3.7 Max                | 340                 | 840                | 1,690            |
-| Qwen3.7 Plus               | 4,300               | 10,800             | 21,600           |
-| Qwen3.6 Plus               | 3,300               | 8,200              | 16,300           |
-| DeepSeek V4 Pro            | 1,050               | 2,600              | 5,200            |
-| DeepSeek V4 Flash          | 7,600               | 18,900             | 37,800           |
-| Hy3                        | 4,300               | 10,750             | 21,500           |
-| Ox Alpha Free              | -                   | -                  | -                |
+| Model                        | الطلبات لكل 5 ساعات | الطلبات في الأسبوع | الطلبات في الشهر |
+| ---------------------------- | ------------------- | ------------------ | ---------------- |
+| Grok 4.5                     | 120                 | 300                | 600              |
+| GPT 5.6 Luna                 | 2,050               | 5,100              | 10,250           |
+| GLM-5.3                      | 220                 | 540                | 1,080            |
+| GLM-5.2                      | 880                 | 2,150              | 4,300            |
+| GLM-5.1                      | 880                 | 2,150              | 4,300            |
+| Kimi K3                      | 110                 | 250                | 490              |
+| Kimi K2.7 Code               | 1,350               | 3,380              | 6,750            |
+| Kimi K2.6                    | 1,150               | 2,880              | 5,750            |
+| MiMo-V2.5                    | 30,100              | 75,200             | 150,400          |
+| MiMo-V2.5-Pro                | 3,250               | 8,150              | 16,300           |
+| MiniMax M3                   | 3,200               | 8,000              | 16,000           |
+| MiniMax M2.7                 | 3,400               | 8,500              | 17,000           |
+| Muse Spark 1.2 Contributor   | 45,300              | 113,300            | 226,600          |
+| Qwen3.8 Max                  | 160                 | 400                | 810              |
+| Qwen3.7 Max                  | 340                 | 840                | 1,690            |
+| Qwen3.7 Plus                 | 4,300               | 10,800             | 21,600           |
+| Qwen3.6 Plus                 | 3,300               | 8,200              | 16,300           |
+| DeepSeek V4 Pro              | 1,050               | 2,600              | 5,200            |
+| DeepSeek V4 Flash            | 7,600               | 18,900             | 37,800           |
+| DeepSeek V4 Flash Vision Exp | 3,800               | 9,450              | 18,900           |
+| Hy3                          | 4,300               | 10,750             | 21,500           |
+| Ox Alpha Free                | -                   | -                  | -                |

 تستند التقديرات إلى أنماط الطلبات المرصودة:

@@ -120,6 +122,7 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال
 - Kimi K2.7/K2.6 — ‏870 input، و55,000 cached، و200 output tokens لكل طلب
 - DeepSeek V4 Pro — ‏750 input، و82,000 cached، و290 output tokens لكل طلب
 - DeepSeek V4 Flash — ‏410 input، و71,300 cached، و310 output tokens لكل طلب
+- DeepSeek V4 Flash Vision Exp — ‏410 input، و71,300 cached، و310 output tokens لكل طلب
 - MiniMax M3 — ‏510 input، و56,000 cached، و190 output tokens لكل طلب
 - MiniMax M2.7 — ‏300 input، و55,000 cached، و125 output tokens لكل طلب
 - Muse Spark 1.2 Contributor — ‏620 input، و71,400 cached، و300 output tokens لكل طلب
@@ -133,37 +136,41 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال

 تستند التقديرات أيضًا إلى الأسعار التالية لكل 1M tokens والاستخدام الشهري المتضمن مع كل نموذج:

-| النموذج                      | الإدخال | الإخراج | القراءة المخزنة | الكتابة المخزنة | الاستخدام |
-| ---------------------------- | ------- | ------- | --------------- | --------------- | --------- |
-| Grok 4.5                     | $2.00   | $6.00   | $0.30           | -               | $15       |
-| GPT 5.6 Luna (≤ 272K tokens) | $0.20   | $1.20   | $0.02           | $0.25           | $15       |
-| GPT 5.6 Luna (> 272K tokens) | $0.40   | $1.80   | $0.04           | $0.50           | $15       |
-| GLM-5.3                      | $1.40   | $4.40   | $0.26           | -               | $15       |
-| GLM-5.2                      | $1.40   | $4.40   | $0.26           | -               | $60       |
-| GLM-5.1                      | $1.40   | $4.40   | $0.26           | -               | $60       |
-| Kimi K3                      | $3.00   | $15.00  | $0.30           | -               | $15       |
-| Kimi K2.7 Code               | $0.95   | $4.00   | $0.19           | -               | $60       |
-| Kimi K2.6                    | $0.95   | $4.00   | $0.16           | -               | $60       |
-| MiMo V2.5                    | $0.14   | $0.28   | $0.0028         | -               | $60       |
-| MiMo V2.5 Pro                | $0.435  | $0.87   | $0.003625       | -               | $15       |
-| MiniMax M3                   | $0.30   | $1.20   | $0.06           | -               | $60       |
-| MiniMax M2.7                 | $0.30   | $1.20   | $0.06           | $0.375          | $60       |
-| MiniMax M2.5                 | $0.30   | $1.20   | $0.06           | $0.375          | $60       |
-| Muse Spark 1.2 Contributor   | $0.10   | $0.20   | $0.002          | -               | $60       |
-| Qwen3.8 Max                  | $2.00   | $6.00   | $0.25           | $2.50           | $15       |
-| Qwen3.7 Max                  | $2.50   | $7.50   | $0.50           | $3.125          | $60       |
-| Qwen3.7 Plus (≤ 256K tokens) | $0.40   | $1.60   | $0.04           | $0.50           | $60       |
-| Qwen3.7 Plus (> 256K tokens) | $1.20   | $4.80   | $0.12           | $1.50           | $60       |
-| Qwen3.6 Plus (≤ 256K tokens) | $0.50   | $3.00   | $0.05           | $0.625          | $60       |
-| Qwen3.6 Plus (> 256K tokens) | $2.00   | $6.00   | $0.20           | $2.50           | $60       |
-| DeepSeek V4 Pro (Off-Peak)   | $0.66   | $1.98   | $0.022          | -               | $15       |
-| DeepSeek V4 Pro (Peak)       | $1.32   | $3.96   | $0.044          | -               | $15       |
-| DeepSeek V4 Flash (Off-Peak) | $0.22   | $0.66   | $0.007          | -               | $30       |
-| DeepSeek V4 Flash (Peak)     | $0.44   | $1.32   | $0.014          | -               | $30       |
-| Hy3                          | $0.14   | $0.58   | $0.035          | -               | $60       |
-| Ox Alpha Free                | -       | -       | -               | -               | -         |
+| النموذج                                 | الإدخال | الإخراج | القراءة المخزنة | الكتابة المخزنة | الاستخدام |
+| --------------------------------------- | ------- | ------- | --------------- | --------------- | --------- |
+| Grok 4.5                                | $2.00   | $6.00   | $0.30           | -               | $15       |
+| GPT 5.6 Luna (≤ 272K tokens)            | $0.20   | $1.20   | $0.02           | $0.25           | $15       |
+| GPT 5.6 Luna (> 272K tokens)            | $0.40   | $1.80   | $0.04           | $0.50           | $15       |
+| GLM-5.3                                 | $1.40   | $4.40   | $0.26           | -               | $15       |
+| GLM-5.2                                 | $1.40   | $4.40   | $0.26           | -               | $60       |
+| GLM-5.1                                 | $1.40   | $4.40   | $0.26           | -               | $60       |
+| Kimi K3                                 | $3.00   | $15.00  | $0.30           | -               | $15       |
+| Kimi K2.7 Code                          | $0.95   | $4.00   | $0.19           | -               | $60       |
+| Kimi K2.6                               | $0.95   | $4.00   | $0.16           | -               | $60       |
+| MiMo V2.5                               | $0.14   | $0.28   | $0.0028         | -               | $60       |
+| MiMo V2.5 Pro                           | $0.435  | $0.87   | $0.003625       | -               | $15       |
+| MiniMax M3                              | $0.30   | $1.20   | $0.06           | -               | $60       |
+| MiniMax M2.7                            | $0.30   | $1.20   | $0.06           | $0.375          | $60       |
+| MiniMax M2.5                            | $0.30   | $1.20   | $0.06           | $0.375          | $60       |
+| Muse Spark 1.2 Contributor              | $0.10   | $0.20   | $0.002          | -               | $60       |
+| Qwen3.8 Max                             | $2.00   | $6.00   | $0.25           | $2.50           | $15       |
+| Qwen3.7 Max                             | $2.50   | $7.50   | $0.50           | $3.125          | $60       |
+| Qwen3.7 Plus (≤ 256K tokens)            | $0.40   | $1.60   | $0.04           | $0.50           | $60       |
+| Qwen3.7 Plus (> 256K tokens)            | $1.20   | $4.80   | $0.12           | $1.50           | $60       |
+| Qwen3.6 Plus (≤ 256K tokens)            | $0.50   | $3.00   | $0.05           | $0.625          | $60       |
+| Qwen3.6 Plus (> 256K tokens)            | $2.00   | $6.00   | $0.20           | $2.50           | $60       |
+| DeepSeek V4 Pro (Off-Peak)              | $0.66   | $1.98   | $0.022          | -               | $15       |
+| DeepSeek V4 Pro (Peak)                  | $1.32   | $3.96   | $0.044          | -               | $15       |
+| DeepSeek V4 Flash (Off-Peak)            | $0.22   | $0.66   | $0.007          | -               | $30       |
+| DeepSeek V4 Flash (Peak)                | $0.44   | $1.32   | $0.014          | -               | $30       |
+| DeepSeek V4 Flash Vision Exp (Off-Peak) | $0.22   | $0.66   | $0.007          | -               | $15       |
+| DeepSeek V4 Flash Vision Exp (Peak)     | $0.44   | $1.32   | $0.014          | -               | $15       |
+| Hy3                                     | $0.14   | $0.58   | $0.035          | -               | $60       |
+| Ox Alpha Free                           | -       | -       | -               | -               | -         |

-**DeepSeek V4 Flash / Pro:** ساعات Peak هي 01:00-04:00 و06:00-10:00 UTC؛ وجميع الساعات الأخرى Off-Peak. [اعرف المزيد](https://api-docs.deepseek.com/quick_start/pricing/).
+**DeepSeek V4 Flash / V4 Flash Vision Exp / V4 Pro:** ساعات Peak هي 01:00-04:00 و06:00-10:00 UTC؛ وجميع الساعات الأخرى Off-Peak. [اعرف المزيد](https://api-docs.deepseek.com/quick_start/pricing/).
+
+**DeepSeek V4 Flash Vision Exp:** يتم تحويل الصور إلى رموز بناءً على أبعادها، وتُحتسب كرموز إدخال إلى جانب رموز النص. [اعرف المزيد](https://api-docs.deepseek.com/quick_start/pricing/).

 **Ox Alpha Free:** مجاني لفترة محدودة.

@@ -199,30 +206,31 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال

 يمكنك أيضًا الوصول إلى نماذج Go عبر نقاط نهاية API التالية.

-| Model                      | Model ID                   | Endpoint                                         | AI SDK Package              |
-| -------------------------- | -------------------------- | ------------------------------------------------ | --------------------------- |
-| Grok 4.5                   | grok-4.5                   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GPT 5.6 Luna               | gpt-5.6-luna               | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3                    | glm-5.3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.2                    | glm-5.2                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.1                    | glm-5.1                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K3                    | kimi-k3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code             | kimi-k2.7-code             | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.6                  | kimi-k2.6                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Pro            | deepseek-v4-pro            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash          | deepseek-v4-flash          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5                  | mimo-v2.5                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro              | mimo-v2.5-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiniMax M3                 | minimax-m3                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.7               | minimax-m2.7               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.5               | minimax-m2.5               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Muse Spark 1.2 Contributor | muse-spark-1.2-contributor | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| Qwen3.8 Max                | qwen3.8-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Max                | qwen3.7-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus               | qwen3.7-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus               | qwen3.6-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Hy3                        | hy3                        | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Ox Alpha Free              | x-preview-f-free           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Model                        | Model ID                     | Endpoint                                         | AI SDK Package              |
+| ---------------------------- | ---------------------------- | ------------------------------------------------ | --------------------------- |
+| Grok 4.5                     | grok-4.5                     | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                 | gpt-5.6-luna                 | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GLM-5.3                      | glm-5.3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.2                      | glm-5.2                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.1                      | glm-5.1                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K3                      | kimi-k3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code               | kimi-k2.7-code               | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                    | kimi-k2.6                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Pro              | deepseek-v4-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash            | deepseek-v4-flash            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Vision Exp | deepseek-v4-flash-vision-exp | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5                    | mimo-v2.5                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro                | mimo-v2.5-pro                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiniMax M3                   | minimax-m3                   | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.7                 | minimax-m2.7                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.5                 | minimax-m2.5                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Muse Spark 1.2 Contributor   | muse-spark-1.2-contributor   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| Qwen3.8 Max                  | qwen3.8-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Max                  | qwen3.7-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                 | qwen3.7-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                 | qwen3.6-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Hy3                          | hy3                          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Ox Alpha Free                | ox-alpha-free                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |

 يستخدم [model id](/docs/config/#models) في إعدادات OpenCode لديك التنسيق `opencode-go/<model-id>`. على سبيل المثال، بالنسبة إلى Kimi K3، ستستخدم `opencode-go/kimi-k3` في إعداداتك.

@@ -240,29 +248,30 @@ https://opencode.ai/zen/go/v1/models

 ## الخصوصية

-| النموذج                    | تدريب النموذج | الاحتفاظ بالبيانات |
-| -------------------------- | ------------- | ------------------ |
-| Grok 4.5                   | غير مستخدَمة  | 30 يومًا           |
-| GPT 5.6 Luna               | غير مستخدَمة  | 30 يومًا           |
-| GLM-5.3                    | غير مستخدَمة  | 0 أيام             |
-| GLM-5.2                    | غير مستخدَمة  | 0 أيام             |
-| GLM-5.1                    | غير مستخدَمة  | 0 أيام             |
-| Kimi K3                    | غير مستخدَمة  | 0 أيام             |
-| Kimi K2.7 Code             | غير مستخدَمة  | 0 أيام             |
-| Kimi K2.6                  | غير مستخدَمة  | 0 أيام             |
-| MiMo-V2.5-Pro              | غير مستخدَمة  | 0 أيام             |
-| MiMo-V2.5                  | غير مستخدَمة  | 0 أيام             |
-| Qwen3.8 Max                | غير مستخدَمة  | 0 أيام             |
-| Qwen3.7 Max                | غير مستخدَمة  | 0 أيام             |
-| Qwen3.7 Plus               | غير مستخدَمة  | 0 أيام             |
-| Qwen3.6 Plus               | غير مستخدَمة  | 0 أيام             |
-| MiniMax M3                 | غير مستخدَمة  | 0 أيام             |
-| MiniMax M2.7               | غير مستخدَمة  | 0 أيام             |
-| Muse Spark 1.2 Contributor | نعم           | ليست ZDR           |
-| DeepSeek V4 Pro            | غير مستخدَمة  | 0 أيام             |
-| DeepSeek V4 Flash          | غير مستخدَمة  | 0 أيام             |
-| Hy3                        | غير مستخدَمة  | 0 أيام             |
-| Ox Alpha Free              | غير مستخدَمة  | 0 أيام             |
+| النموذج                      | تدريب النموذج | الاحتفاظ بالبيانات |
+| ---------------------------- | ------------- | ------------------ |
+| Grok 4.5                     | غير مستخدَمة  | 30 يومًا           |
+| GPT 5.6 Luna                 | غير مستخدَمة  | 30 يومًا           |
+| GLM-5.3                      | غير مستخدَمة  | 0 أيام             |
+| GLM-5.2                      | غير مستخدَمة  | 0 أيام             |
+| GLM-5.1                      | غير مستخدَمة  | 0 أيام             |
+| Kimi K3                      | غير مستخدَمة  | 0 أيام             |
+| Kimi K2.7 Code               | غير مستخدَمة  | 0 أيام             |
+| Kimi K2.6                    | غير مستخدَمة  | 0 أيام             |
+| MiMo-V2.5-Pro                | غير مستخدَمة  | 0 أيام             |
+| MiMo-V2.5                    | غير مستخدَمة  | 0 أيام             |
+| Qwen3.8 Max                  | غير مستخدَمة  | 0 أيام             |
+| Qwen3.7 Max                  | غير مستخدَمة  | 0 أيام             |
+| Qwen3.7 Plus                 | غير مستخدَمة  | 0 أيام             |
+| Qwen3.6 Plus                 | غير مستخدَمة  | 0 أيام             |
+| MiniMax M3                   | غير مستخدَمة  | 0 أيام             |
+| MiniMax M2.7                 | غير مستخدَمة  | 0 أيام             |
+| Muse Spark 1.2 Contributor   | نعم           | ليست ZDR           |
+| DeepSeek V4 Pro              | غير مستخدَمة  | 0 أيام             |
+| DeepSeek V4 Flash            | غير مستخدَمة  | 0 أيام             |
+| DeepSeek V4 Flash Vision Exp | غير مستخدَمة  | 0 أيام             |
+| Hy3                          | غير مستخدَمة  | 0 أيام             |
+| Ox Alpha Free                | غير مستخدَمة  | 0 أيام             |

 - **Grok 4.5:** تعطّل ZDR ميزات API مهمة تعتمد على البيانات المخزنة، بما في ذلك Responses API ذات الحالة، وFiles and Collections، وBatch API. [اعرف المزيد](https://docs.x.ai/developers/faq/security#what-is-zero-data-retention-zdr).
 - **GPT 5.6 Luna:** تُنشأ سجلات مراقبة إساءة الاستخدام لكل استخدام لميزات API، ويُحتفظ بها لمدة تصل إلى 30 يومًا. [اعرف المزيد](https://developers.openai.com/api/docs/guides/your-data#data-retention-controls-for-abuse-monitoring).
@@ -188,8 +188,8 @@ https://opencode.ai/zen/v1/models
 | Grok 4.5 (> 200K tokens)          | $4.00   | $12.00  | $0.60           | -               |
 | Grok Build 0.1                    | $1.00   | $2.00   | $0.20           | -               |
 | Muse Spark 1.2                    | $1.25   | $4.25   | $0.15           | -               |
-| GPT 5.6 Sol (≤ 272K tokens)       | $5.00   | $30.00  | $0.50           | $6.25           |
-| GPT 5.6 Sol (> 272K tokens)       | $10.00  | $45.00  | $1.00           | $12.50          |
+| GPT 5.6 Sol (≤ 272K tokens)       | $2.00   | $10.00  | $0.20           | $2.50           |
+| GPT 5.6 Sol (> 272K tokens)       | $4.00   | $15.00  | $0.40           | $5.00           |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00   | $12.00  | $0.20           | $2.50           |
 | GPT 5.6 Terra (> 272K tokens)     | $4.00   | $18.00  | $0.40           | $5.00           |
 | GPT 5.6 Luna (≤ 272K tokens)      | $0.20   | $1.20   | $0.02           | $0.25           |
@@ -214,6 +214,8 @@ https://opencode.ai/zen/v1/models
 | GPT 5 Codex                       | $1.07   | $8.50   | $0.107          | -               |
 | GPT 5 Nano                        | $0.05   | $0.40   | $0.005          | -               |

+**GPT 5.6 Sol:** تشمل الأسعار المعروضة خصمًا بنسبة 50% حتى 18 سبتمبر 2026.
+
 **DeepSeek V4 Flash / Pro:** ساعات Peak هي 01:00-04:00 و06:00-10:00 UTC؛ وجميع الساعات الأخرى Off-Peak. [اعرف المزيد](https://api-docs.deepseek.com/quick_start/pricing/).

 قد تلاحظ [نماذج منخفضة التكلفة](/docs/config/#models)، مثل Haiku أو Nano أو Flash، في سجل الاستخدام. يستخدم OpenCode هذه النماذج لإنشاء عناوين الجلسات.
@@ -78,6 +78,7 @@ Trenutna lista modela uključuje:
 - **Qwen3.6 Plus**
 - **DeepSeek V4 Pro**
 - **DeepSeek V4 Flash**
+- **DeepSeek V4 Flash Vision Exp**
 - **Hy3**
 - **Ox Alpha Free** (ograničeno vrijeme)

@@ -97,29 +98,30 @@ Ograničenja su definisana u dolarskoj vrijednosti. To znači da vaš stvarni br

 Tabela ispod pruža procijenjeni broj zahtjeva na osnovu tipičnih obrazaca korištenja Go pretplate:

-| Model                      | zahtjeva na 5 sati | zahtjeva sedmično | zahtjeva mjesečno |
-| -------------------------- | ------------------ | ----------------- | ----------------- |
-| Grok 4.5                   | 120                | 300               | 600               |
-| GPT 5.6 Luna               | 2,050              | 5,100             | 10,250            |
-| GLM-5.3                    | 220                | 540               | 1,080             |
-| GLM-5.2                    | 880                | 2,150             | 4,300             |
-| GLM-5.1                    | 880                | 2,150             | 4,300             |
-| Kimi K3                    | 110                | 250               | 490               |
-| Kimi K2.7 Code             | 1,350              | 3,380             | 6,750             |
-| Kimi K2.6                  | 1,150              | 2,880             | 5,750             |
-| MiMo-V2.5                  | 30,100             | 75,200            | 150,400           |
-| MiMo-V2.5-Pro              | 3,250              | 8,150             | 16,300            |
-| MiniMax M3                 | 3,200              | 8,000             | 16,000            |
-| MiniMax M2.7               | 3,400              | 8,500             | 17,000            |
-| Muse Spark 1.2 Contributor | 45,300             | 113,300           | 226,600           |
-| Qwen3.8 Max                | 160                | 400               | 810               |
-| Qwen3.7 Max                | 340                | 840               | 1,690             |
-| Qwen3.7 Plus               | 4,300              | 10,800            | 21,600            |
-| Qwen3.6 Plus               | 3,300              | 8,200             | 16,300            |
-| DeepSeek V4 Pro            | 1,050              | 2,600             | 5,200             |
-| DeepSeek V4 Flash          | 7,600              | 18,900            | 37,800            |
-| Hy3                        | 4,300              | 10,750            | 21,500            |
-| Ox Alpha Free              | -                  | -                 | -                 |
+| Model                        | zahtjeva na 5 sati | zahtjeva sedmično | zahtjeva mjesečno |
+| ---------------------------- | ------------------ | ----------------- | ----------------- |
+| Grok 4.5                     | 120                | 300               | 600               |
+| GPT 5.6 Luna                 | 2,050              | 5,100             | 10,250            |
+| GLM-5.3                      | 220                | 540               | 1,080             |
+| GLM-5.2                      | 880                | 2,150             | 4,300             |
+| GLM-5.1                      | 880                | 2,150             | 4,300             |
+| Kimi K3                      | 110                | 250               | 490               |
+| Kimi K2.7 Code               | 1,350              | 3,380             | 6,750             |
+| Kimi K2.6                    | 1,150              | 2,880             | 5,750             |
+| MiMo-V2.5                    | 30,100             | 75,200            | 150,400           |
+| MiMo-V2.5-Pro                | 3,250              | 8,150             | 16,300            |
+| MiniMax M3                   | 3,200              | 8,000             | 16,000            |
+| MiniMax M2.7                 | 3,400              | 8,500             | 17,000            |
+| Muse Spark 1.2 Contributor   | 45,300             | 113,300           | 226,600           |
+| Qwen3.8 Max                  | 160                | 400               | 810               |
+| Qwen3.7 Max                  | 340                | 840               | 1,690             |
+| Qwen3.7 Plus                 | 4,300              | 10,800            | 21,600            |
+| Qwen3.6 Plus                 | 3,300              | 8,200             | 16,300            |
+| DeepSeek V4 Pro              | 1,050              | 2,600             | 5,200             |
+| DeepSeek V4 Flash            | 7,600              | 18,900            | 37,800            |
+| DeepSeek V4 Flash Vision Exp | 3,800              | 9,450             | 18,900            |
+| Hy3                          | 4,300              | 10,750            | 21,500            |
+| Ox Alpha Free                | -                  | -                 | -                 |

 Procjene se zasnivaju na zapaženim obrascima zahtjeva:

@@ -130,6 +132,7 @@ Procjene se zasnivaju na zapaženim obrascima zahtjeva:
 - Kimi K2.7/K2.6 — 870 ulaznih, 55,000 keširanih, 200 izlaznih tokena po zahtjevu
 - DeepSeek V4 Pro — 750 ulaznih, 82,000 keširanih, 290 izlaznih tokena po zahtjevu
 - DeepSeek V4 Flash — 410 ulaznih, 71,300 keširanih, 310 izlaznih tokena po zahtjevu
+- DeepSeek V4 Flash Vision Exp — 410 ulaznih, 71,300 keširanih, 310 izlaznih tokena po zahtjevu
 - MiniMax M3 — 510 ulaznih, 56,000 keširanih, 190 izlaznih tokena po zahtjevu
 - MiniMax M2.7 — 300 ulaznih, 55,000 keširanih, 125 izlaznih tokena po zahtjevu
 - Muse Spark 1.2 Contributor — 620 ulaznih, 71,400 keširanih, 300 izlaznih tokena po zahtjevu
@@ -143,37 +146,41 @@ Procjene se zasnivaju na zapaženim obrascima zahtjeva:

 Procjene se također zasnivaju na sljedećim cijenama po 1M tokena i mjesečnoj potrošnji uključenoj uz svaki model:

-| Model                        | Input  | Output | Cached Read | Cached Write | Potrošnja |
-| ---------------------------- | ------ | ------ | ----------- | ------------ | --------- |
-| Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15       |
-| GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15       |
-| GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15       |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15       |
-| GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60       |
-| GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60       |
-| Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15       |
-| Kimi K2.7 Code               | $0.95  | $4.00  | $0.19       | -            | $60       |
-| Kimi K2.6                    | $0.95  | $4.00  | $0.16       | -            | $60       |
-| MiMo V2.5                    | $0.14  | $0.28  | $0.0028     | -            | $60       |
-| MiMo V2.5 Pro                | $0.435 | $0.87  | $0.003625   | -            | $15       |
-| MiniMax M3                   | $0.30  | $1.20  | $0.06       | -            | $60       |
-| MiniMax M2.7                 | $0.30  | $1.20  | $0.06       | $0.375       | $60       |
-| MiniMax M2.5                 | $0.30  | $1.20  | $0.06       | $0.375       | $60       |
-| Muse Spark 1.2 Contributor   | $0.10  | $0.20  | $0.002      | -            | $60       |
-| Qwen3.8 Max                  | $2.00  | $6.00  | $0.25       | $2.50        | $15       |
-| Qwen3.7 Max                  | $2.50  | $7.50  | $0.50       | $3.125       | $60       |
-| Qwen3.7 Plus (≤ 256K tokens) | $0.40  | $1.60  | $0.04       | $0.50        | $60       |
-| Qwen3.7 Plus (> 256K tokens) | $1.20  | $4.80  | $0.12       | $1.50        | $60       |
-| Qwen3.6 Plus (≤ 256K tokens) | $0.50  | $3.00  | $0.05       | $0.625       | $60       |
-| Qwen3.6 Plus (> 256K tokens) | $2.00  | $6.00  | $0.20       | $2.50        | $60       |
-| DeepSeek V4 Pro (Off-Peak)   | $0.66  | $1.98  | $0.022      | -            | $15       |
-| DeepSeek V4 Pro (Peak)       | $1.32  | $3.96  | $0.044      | -            | $15       |
-| DeepSeek V4 Flash (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $30       |
-| DeepSeek V4 Flash (Peak)     | $0.44  | $1.32  | $0.014      | -            | $30       |
-| Hy3                          | $0.14  | $0.58  | $0.035      | -            | $60       |
-| Ox Alpha Free                | -      | -      | -           | -            | -         |
+| Model                                   | Input  | Output | Cached Read | Cached Write | Potrošnja |
+| --------------------------------------- | ------ | ------ | ----------- | ------------ | --------- |
+| Grok 4.5                                | $2.00  | $6.00  | $0.30       | -            | $15       |
+| GPT 5.6 Luna (≤ 272K tokens)            | $0.20  | $1.20  | $0.02       | $0.25        | $15       |
+| GPT 5.6 Luna (> 272K tokens)            | $0.40  | $1.80  | $0.04       | $0.50        | $15       |
+| GLM-5.3                                 | $1.40  | $4.40  | $0.26       | -            | $15       |
+| GLM-5.2                                 | $1.40  | $4.40  | $0.26       | -            | $60       |
+| GLM-5.1                                 | $1.40  | $4.40  | $0.26       | -            | $60       |
+| Kimi K3                                 | $3.00  | $15.00 | $0.30       | -            | $15       |
+| Kimi K2.7 Code                          | $0.95  | $4.00  | $0.19       | -            | $60       |
+| Kimi K2.6                               | $0.95  | $4.00  | $0.16       | -            | $60       |
+| MiMo V2.5                               | $0.14  | $0.28  | $0.0028     | -            | $60       |
+| MiMo V2.5 Pro                           | $0.435 | $0.87  | $0.003625   | -            | $15       |
+| MiniMax M3                              | $0.30  | $1.20  | $0.06       | -            | $60       |
+| MiniMax M2.7                            | $0.30  | $1.20  | $0.06       | $0.375       | $60       |
+| MiniMax M2.5                            | $0.30  | $1.20  | $0.06       | $0.375       | $60       |
+| Muse Spark 1.2 Contributor              | $0.10  | $0.20  | $0.002      | -            | $60       |
+| Qwen3.8 Max                             | $2.00  | $6.00  | $0.25       | $2.50        | $15       |
+| Qwen3.7 Max                             | $2.50  | $7.50  | $0.50       | $3.125       | $60       |
+| Qwen3.7 Plus (≤ 256K tokens)            | $0.40  | $1.60  | $0.04       | $0.50        | $60       |
+| Qwen3.7 Plus (> 256K tokens)            | $1.20  | $4.80  | $0.12       | $1.50        | $60       |
+| Qwen3.6 Plus (≤ 256K tokens)            | $0.50  | $3.00  | $0.05       | $0.625       | $60       |
+| Qwen3.6 Plus (> 256K tokens)            | $2.00  | $6.00  | $0.20       | $2.50        | $60       |
+| DeepSeek V4 Pro (Off-Peak)              | $0.66  | $1.98  | $0.022      | -            | $15       |
+| DeepSeek V4 Pro (Peak)                  | $1.32  | $3.96  | $0.044      | -            | $15       |
+| DeepSeek V4 Flash (Off-Peak)            | $0.22  | $0.66  | $0.007      | -            | $30       |
+| DeepSeek V4 Flash (Peak)                | $0.44  | $1.32  | $0.014      | -            | $30       |
+| DeepSeek V4 Flash Vision Exp (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $15       |
+| DeepSeek V4 Flash Vision Exp (Peak)     | $0.44  | $1.32  | $0.014      | -            | $15       |
+| Hy3                                     | $0.14  | $0.58  | $0.035      | -            | $60       |
+| Ox Alpha Free                           | -      | -      | -           | -            | -         |

-**DeepSeek V4 Flash / Pro:** Peak sati su 01:00-04:00 i 06:00-10:00 UTC; svi ostali sati su Off-Peak. [Saznajte više](https://api-docs.deepseek.com/quick_start/pricing/).
+**DeepSeek V4 Flash / V4 Flash Vision Exp / V4 Pro:** Peak sati su 01:00-04:00 i 06:00-10:00 UTC; svi ostali sati su Off-Peak. [Saznajte više](https://api-docs.deepseek.com/quick_start/pricing/).
+
+**DeepSeek V4 Flash Vision Exp:** Slike se pretvaraju u tokene na osnovu svojih dimenzija i naplaćuju kao ulazni tokeni zajedno s tekstualnim tokenima. [Saznajte više](https://api-docs.deepseek.com/quick_start/pricing/).

 **Ox Alpha Free:** Besplatan ograničeno vrijeme.

@@ -211,30 +218,31 @@ Za ove modele i dalje dobijate malo više nego da direktno plaćate provajderima

 Također možete pristupiti Go modelima putem sljedećih API endpointa.

-| Model                      | Model ID                   | Endpoint                                         | AI SDK Paket                |
-| -------------------------- | -------------------------- | ------------------------------------------------ | --------------------------- |
-| Grok 4.5                   | grok-4.5                   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GPT 5.6 Luna               | gpt-5.6-luna               | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3                    | glm-5.3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.2                    | glm-5.2                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.1                    | glm-5.1                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K3                    | kimi-k3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code             | kimi-k2.7-code             | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.6                  | kimi-k2.6                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Pro            | deepseek-v4-pro            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash          | deepseek-v4-flash          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5                  | mimo-v2.5                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro              | mimo-v2.5-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiniMax M3                 | minimax-m3                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.7               | minimax-m2.7               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.5               | minimax-m2.5               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Muse Spark 1.2 Contributor | muse-spark-1.2-contributor | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| Qwen3.8 Max                | qwen3.8-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Max                | qwen3.7-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus               | qwen3.7-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus               | qwen3.6-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Hy3                        | hy3                        | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Ox Alpha Free              | x-preview-f-free           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Model                        | Model ID                     | Endpoint                                         | AI SDK Paket                |
+| ---------------------------- | ---------------------------- | ------------------------------------------------ | --------------------------- |
+| Grok 4.5                     | grok-4.5                     | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                 | gpt-5.6-luna                 | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GLM-5.3                      | glm-5.3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.2                      | glm-5.2                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.1                      | glm-5.1                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K3                      | kimi-k3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code               | kimi-k2.7-code               | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                    | kimi-k2.6                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Pro              | deepseek-v4-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash            | deepseek-v4-flash            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Vision Exp | deepseek-v4-flash-vision-exp | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5                    | mimo-v2.5                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro                | mimo-v2.5-pro                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiniMax M3                   | minimax-m3                   | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.7                 | minimax-m2.7                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.5                 | minimax-m2.5                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Muse Spark 1.2 Contributor   | muse-spark-1.2-contributor   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| Qwen3.8 Max                  | qwen3.8-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Max                  | qwen3.7-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                 | qwen3.7-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                 | qwen3.6-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Hy3                          | hy3                          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Ox Alpha Free                | ox-alpha-free                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |

 [Model id](/docs/config/#models) u vašoj OpenCode konfiguraciji
 koristi format `opencode-go/<model-id>`. Na primjer, za Kimi K3, koristili biste
@@ -254,29 +262,30 @@ https://opencode.ai/zen/go/v1/models

 ## Privatnost

-| Model                      | Treniranje modela | Zadržavanje podataka |
-| -------------------------- | ----------------- | -------------------- |
-| Grok 4.5                   | Ne koristi se     | 30 dana              |
-| GPT 5.6 Luna               | Ne koristi se     | 30 dana              |
-| GLM-5.3                    | Ne koristi se     | 0 dana               |
-| GLM-5.2                    | Ne koristi se     | 0 dana               |
-| GLM-5.1                    | Ne koristi se     | 0 dana               |
-| Kimi K3                    | Ne koristi se     | 0 dana               |
-| Kimi K2.7 Code             | Ne koristi se     | 0 dana               |
-| Kimi K2.6                  | Ne koristi se     | 0 dana               |
-| MiMo-V2.5-Pro              | Ne koristi se     | 0 dana               |
-| MiMo-V2.5                  | Ne koristi se     | 0 dana               |
-| Qwen3.8 Max                | Ne koristi se     | 0 dana               |
-| Qwen3.7 Max                | Ne koristi se     | 0 dana               |
-| Qwen3.7 Plus               | Ne koristi se     | 0 dana               |
-| Qwen3.6 Plus               | Ne koristi se     | 0 dana               |
-| MiniMax M3                 | Ne koristi se     | 0 dana               |
-| MiniMax M2.7               | Ne koristi se     | 0 dana               |
-| Muse Spark 1.2 Contributor | Da                | Nije ZDR             |
-| DeepSeek V4 Pro            | Ne koristi se     | 0 dana               |
-| DeepSeek V4 Flash          | Ne koristi se     | 0 dana               |
-| Hy3                        | Ne koristi se     | 0 dana               |
-| Ox Alpha Free              | Ne koristi se     | 0 dana               |
+| Model                        | Treniranje modela | Zadržavanje podataka |
+| ---------------------------- | ----------------- | -------------------- |
+| Grok 4.5                     | Ne koristi se     | 30 dana              |
+| GPT 5.6 Luna                 | Ne koristi se     | 30 dana              |
+| GLM-5.3                      | Ne koristi se     | 0 dana               |
+| GLM-5.2                      | Ne koristi se     | 0 dana               |
+| GLM-5.1                      | Ne koristi se     | 0 dana               |
+| Kimi K3                      | Ne koristi se     | 0 dana               |
+| Kimi K2.7 Code               | Ne koristi se     | 0 dana               |
+| Kimi K2.6                    | Ne koristi se     | 0 dana               |
+| MiMo-V2.5-Pro                | Ne koristi se     | 0 dana               |
+| MiMo-V2.5                    | Ne koristi se     | 0 dana               |
+| Qwen3.8 Max                  | Ne koristi se     | 0 dana               |
+| Qwen3.7 Max                  | Ne koristi se     | 0 dana               |
+| Qwen3.7 Plus                 | Ne koristi se     | 0 dana               |
+| Qwen3.6 Plus                 | Ne koristi se     | 0 dana               |
+| MiniMax M3                   | Ne koristi se     | 0 dana               |
+| MiniMax M2.7                 | Ne koristi se     | 0 dana               |
+| Muse Spark 1.2 Contributor   | Da                | Nije ZDR             |
+| DeepSeek V4 Pro              | Ne koristi se     | 0 dana               |
+| DeepSeek V4 Flash            | Ne koristi se     | 0 dana               |
+| DeepSeek V4 Flash Vision Exp | Ne koristi se     | 0 dana               |
+| Hy3                          | Ne koristi se     | 0 dana               |
+| Ox Alpha Free                | Ne koristi se     | 0 dana               |

 - **Grok 4.5:** ZDR onemogućava važne API funkcije koje zavise od pohranjenih podataka, uključujući Responses API s očuvanjem stanja, Files and Collections i Batch API. [Saznajte više](https://docs.x.ai/developers/faq/security#what-is-zero-data-retention-zdr).
 - **GPT 5.6 Luna:** Zapisi o nadzoru zloupotrebe generišu se za svako korištenje API funkcija i čuvaju do 30 dana. [Saznajte više](https://developers.openai.com/api/docs/guides/your-data#data-retention-controls-for-abuse-monitoring).
@@ -195,8 +195,8 @@ Podržavamo pay-as-you-go model. Ispod su cijene **po 1M tokena**.
 | Grok 4.5 (> 200K tokens)          | $4.00  | $12.00  | $0.60       | -            |
 | Grok Build 0.1                    | $1.00  | $2.00   | $0.20       | -            |
 | Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
-| GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
-| GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
+| GPT 5.6 Sol (≤ 272K tokens)       | $2.00  | $10.00  | $0.20       | $2.50        |
+| GPT 5.6 Sol (> 272K tokens)       | $4.00  | $15.00  | $0.40       | $5.00        |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
 | GPT 5.6 Terra (> 272K tokens)     | $4.00  | $18.00  | $0.40       | $5.00        |
 | GPT 5.6 Luna (≤ 272K tokens)      | $0.20  | $1.20   | $0.02       | $0.25        |
@@ -221,6 +221,8 @@ Podržavamo pay-as-you-go model. Ispod su cijene **po 1M tokena**.
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |

+**GPT 5.6 Sol:** Prikazane cijene uključuju popust od 50% do 18. septembra 2026.
+
 **DeepSeek V4 Flash / Pro:** Peak sati su 01:00-04:00 i 06:00-10:00 UTC; svi ostali sati su Off-Peak. [Saznajte više](https://api-docs.deepseek.com/quick_start/pricing/).

 U historiji korištenja možete primijetiti [jeftinije modele](/docs/config/#models), kao što su Haiku, Nano ili Flash. OpenCode koristi ove modele za generisanje naslova sesija.
@@ -78,6 +78,7 @@ Den nuværende liste over modeller inkluderer:
 - **Qwen3.6 Plus**
 - **DeepSeek V4 Pro**
 - **DeepSeek V4 Flash**
+- **DeepSeek V4 Flash Vision Exp**
 - **Hy3**
 - **Ox Alpha Free** (i en begrænset periode)

@@ -97,29 +98,30 @@ Grænserne er defineret i dollarværdi. Det betyder, at dit faktiske antal anmod

 Tabellen nedenfor giver et estimeret antal anmodninger baseret på typiske Go-forbrugsmønstre:

-| Model                      | anmodninger pr. 5 timer | anmodninger pr. uge | anmodninger pr. måned |
-| -------------------------- | ----------------------- | ------------------- | --------------------- |
-| Grok 4.5                   | 120                     | 300                 | 600                   |
-| GPT 5.6 Luna               | 2,050                   | 5,100               | 10,250                |
-| GLM-5.3                    | 220                     | 540                 | 1,080                 |
-| GLM-5.2                    | 880                     | 2,150               | 4,300                 |
-| GLM-5.1                    | 880                     | 2,150               | 4,300                 |
-| Kimi K3                    | 110                     | 250                 | 490                   |
-| Kimi K2.7 Code             | 1,350                   | 3,380               | 6,750                 |
-| Kimi K2.6                  | 1,150                   | 2,880               | 5,750                 |
-| MiMo-V2.5                  | 30,100                  | 75,200              | 150,400               |
-| MiMo-V2.5-Pro              | 3,250                   | 8,150               | 16,300                |
-| MiniMax M3                 | 3,200                   | 8,000               | 16,000                |
-| MiniMax M2.7               | 3,400                   | 8,500               | 17,000                |
-| Muse Spark 1.2 Contributor | 45,300                  | 113,300             | 226,600               |
-| Qwen3.8 Max                | 160                     | 400                 | 810                   |
-| Qwen3.7 Max                | 340                     | 840                 | 1,690                 |
-| Qwen3.7 Plus               | 4,300                   | 10,800              | 21,600                |
-| Qwen3.6 Plus               | 3,300                   | 8,200               | 16,300                |
-| DeepSeek V4 Pro            | 1,050                   | 2,600               | 5,200                 |
-| DeepSeek V4 Flash          | 7,600                   | 18,900              | 37,800                |
-| Hy3                        | 4,300                   | 10,750              | 21,500                |
-| Ox Alpha Free              | -                       | -                   | -                     |
+| Model                        | anmodninger pr. 5 timer | anmodninger pr. uge | anmodninger pr. måned |
+| ---------------------------- | ----------------------- | ------------------- | --------------------- |
+| Grok 4.5                     | 120                     | 300                 | 600                   |
+| GPT 5.6 Luna                 | 2,050                   | 5,100               | 10,250                |
+| GLM-5.3                      | 220                     | 540                 | 1,080                 |
+| GLM-5.2                      | 880                     | 2,150               | 4,300                 |
+| GLM-5.1                      | 880                     | 2,150               | 4,300                 |
+| Kimi K3                      | 110                     | 250                 | 490                   |
+| Kimi K2.7 Code               | 1,350                   | 3,380               | 6,750                 |
+| Kimi K2.6                    | 1,150                   | 2,880               | 5,750                 |
+| MiMo-V2.5                    | 30,100                  | 75,200              | 150,400               |
+| MiMo-V2.5-Pro                | 3,250                   | 8,150               | 16,300                |
+| MiniMax M3                   | 3,200                   | 8,000               | 16,000                |
+| MiniMax M2.7                 | 3,400                   | 8,500               | 17,000                |
+| Muse Spark 1.2 Contributor   | 45,300                  | 113,300             | 226,600               |
+| Qwen3.8 Max                  | 160                     | 400                 | 810                   |
+| Qwen3.7 Max                  | 340                     | 840                 | 1,690                 |
+| Qwen3.7 Plus                 | 4,300                   | 10,800              | 21,600                |
+| Qwen3.6 Plus                 | 3,300                   | 8,200               | 16,300                |
+| DeepSeek V4 Pro              | 1,050                   | 2,600               | 5,200                 |
+| DeepSeek V4 Flash            | 7,600                   | 18,900              | 37,800                |
+| DeepSeek V4 Flash Vision Exp | 3,800                   | 9,450               | 18,900                |
+| Hy3                          | 4,300                   | 10,750              | 21,500                |
+| Ox Alpha Free                | -                       | -                   | -                     |

 Estimaterne er baseret på observerede anmodningsmønstre:

@@ -130,6 +132,7 @@ Estimaterne er baseret på observerede anmodningsmønstre:
 - Kimi K2.7/K2.6 — 870 input, 55.000 cachelagrede, 200 output-tokens pr. anmodning
 - DeepSeek V4 Pro — 750 input, 82.000 cachelagrede, 290 output-tokens pr. anmodning
 - DeepSeek V4 Flash — 410 input, 71.300 cachelagrede, 310 output-tokens pr. anmodning
+- DeepSeek V4 Flash Vision Exp — 410 input, 71.300 cachelagrede, 310 output-tokens pr. anmodning
 - MiniMax M3 — 510 input, 56.000 cachelagrede, 190 output-tokens pr. anmodning
 - MiniMax M2.7 — 300 input, 55.000 cachelagrede, 125 output-tokens pr. anmodning
 - Muse Spark 1.2 Contributor — 620 input, 71.400 cachelagrede, 300 output-tokens pr. anmodning
@@ -143,37 +146,41 @@ Estimaterne er baseret på observerede anmodningsmønstre:

 Estimaterne er også baseret på følgende priser pr. 1M tokens og det månedlige forbrug, der er inkluderet med hver model:

-| Model                        | Input  | Output | Cached Read | Cached Write | Forbrug |
-| ---------------------------- | ------ | ------ | ----------- | ------------ | ------- |
-| Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15     |
-| GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15     |
-| GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15     |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15     |
-| GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60     |
-| GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60     |
-| Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15     |
-| Kimi K2.7 Code               | $0.95  | $4.00  | $0.19       | -            | $60     |
-| Kimi K2.6                    | $0.95  | $4.00  | $0.16       | -            | $60     |
-| MiMo V2.5                    | $0.14  | $0.28  | $0.0028     | -            | $60     |
-| MiMo V2.5 Pro                | $0.435 | $0.87  | $0.003625   | -            | $15     |
-| MiniMax M3                   | $0.30  | $1.20  | $0.06       | -            | $60     |
-| MiniMax M2.7                 | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
-| MiniMax M2.5                 | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
-| Muse Spark 1.2 Contributor   | $0.10  | $0.20  | $0.002      | -            | $60     |
-| Qwen3.8 Max                  | $2.00  | $6.00  | $0.25       | $2.50        | $15     |
-| Qwen3.7 Max                  | $2.50  | $7.50  | $0.50       | $3.125       | $60     |
-| Qwen3.7 Plus (≤ 256K tokens) | $0.40  | $1.60  | $0.04       | $0.50        | $60     |
-| Qwen3.7 Plus (> 256K tokens) | $1.20  | $4.80  | $0.12       | $1.50        | $60     |
-| Qwen3.6 Plus (≤ 256K tokens) | $0.50  | $3.00  | $0.05       | $0.625       | $60     |
-| Qwen3.6 Plus (> 256K tokens) | $2.00  | $6.00  | $0.20       | $2.50        | $60     |
-| DeepSeek V4 Pro (Off-Peak)   | $0.66  | $1.98  | $0.022      | -            | $15     |
-| DeepSeek V4 Pro (Peak)       | $1.32  | $3.96  | $0.044      | -            | $15     |
-| DeepSeek V4 Flash (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $30     |
-| DeepSeek V4 Flash (Peak)     | $0.44  | $1.32  | $0.014      | -            | $30     |
-| Hy3                          | $0.14  | $0.58  | $0.035      | -            | $60     |
-| Ox Alpha Free                | -      | -      | -           | -            | -       |
+| Model                                   | Input  | Output | Cached Read | Cached Write | Forbrug |
+| --------------------------------------- | ------ | ------ | ----------- | ------------ | ------- |
+| Grok 4.5                                | $2.00  | $6.00  | $0.30       | -            | $15     |
+| GPT 5.6 Luna (≤ 272K tokens)            | $0.20  | $1.20  | $0.02       | $0.25        | $15     |
+| GPT 5.6 Luna (> 272K tokens)            | $0.40  | $1.80  | $0.04       | $0.50        | $15     |
+| GLM-5.3                                 | $1.40  | $4.40  | $0.26       | -            | $15     |
+| GLM-5.2                                 | $1.40  | $4.40  | $0.26       | -            | $60     |
+| GLM-5.1                                 | $1.40  | $4.40  | $0.26       | -            | $60     |
+| Kimi K3                                 | $3.00  | $15.00 | $0.30       | -            | $15     |
+| Kimi K2.7 Code                          | $0.95  | $4.00  | $0.19       | -            | $60     |
+| Kimi K2.6                               | $0.95  | $4.00  | $0.16       | -            | $60     |
+| MiMo V2.5                               | $0.14  | $0.28  | $0.0028     | -            | $60     |
+| MiMo V2.5 Pro                           | $0.435 | $0.87  | $0.003625   | -            | $15     |
+| MiniMax M3                              | $0.30  | $1.20  | $0.06       | -            | $60     |
+| MiniMax M2.7                            | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
+| MiniMax M2.5                            | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
+| Muse Spark 1.2 Contributor              | $0.10  | $0.20  | $0.002      | -            | $60     |
+| Qwen3.8 Max                             | $2.00  | $6.00  | $0.25       | $2.50        | $15     |
+| Qwen3.7 Max                             | $2.50  | $7.50  | $0.50       | $3.125       | $60     |
+| Qwen3.7 Plus (≤ 256K tokens)            | $0.40  | $1.60  | $0.04       | $0.50        | $60     |
+| Qwen3.7 Plus (> 256K tokens)            | $1.20  | $4.80  | $0.12       | $1.50        | $60     |
+| Qwen3.6 Plus (≤ 256K tokens)            | $0.50  | $3.00  | $0.05       | $0.625       | $60     |
+| Qwen3.6 Plus (> 256K tokens)            | $2.00  | $6.00  | $0.20       | $2.50        | $60     |
+| DeepSeek V4 Pro (Off-Peak)              | $0.66  | $1.98  | $0.022      | -            | $15     |
+| DeepSeek V4 Pro (Peak)                  | $1.32  | $3.96  | $0.044      | -            | $15     |
+| DeepSeek V4 Flash (Off-Peak)            | $0.22  | $0.66  | $0.007      | -            | $30     |
+| DeepSeek V4 Flash (Peak)                | $0.44  | $1.32  | $0.014      | -            | $30     |
+| DeepSeek V4 Flash Vision Exp (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $15     |
+| DeepSeek V4 Flash Vision Exp (Peak)     | $0.44  | $1.32  | $0.014      | -            | $15     |
+| Hy3                                     | $0.14  | $0.58  | $0.035      | -            | $60     |
+| Ox Alpha Free                           | -      | -      | -           | -            | -       |

-**DeepSeek V4 Flash / Pro:** Peak-tiderne er 01:00-04:00 og 06:00-10:00 UTC; alle andre tider er Off-Peak. [Læs mere](https://api-docs.deepseek.com/quick_start/pricing/).
+**DeepSeek V4 Flash / V4 Flash Vision Exp / V4 Pro:** Peak-tiderne er 01:00-04:00 og 06:00-10:00 UTC; alle andre tider er Off-Peak. [Læs mere](https://api-docs.deepseek.com/quick_start/pricing/).
+
+**DeepSeek V4 Flash Vision Exp:** Billeder konverteres til tokens baseret på deres dimensioner og afregnes som inputtokens sammen med teksttokens. [Læs mere](https://api-docs.deepseek.com/quick_start/pricing/).

 **Ox Alpha Free:** Gratis i en begrænset periode.

@@ -211,30 +218,31 @@ Med disse modeller får du stadig lidt mere, end hvis du betalte modeludbyderne

 Du kan også få adgang til Go-modeller gennem følgende API-endpoints.

-| Model                      | Model ID                   | Endpoint                                         | AI SDK Package              |
-| -------------------------- | -------------------------- | ------------------------------------------------ | --------------------------- |
-| Grok 4.5                   | grok-4.5                   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GPT 5.6 Luna               | gpt-5.6-luna               | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3                    | glm-5.3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.2                    | glm-5.2                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.1                    | glm-5.1                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K3                    | kimi-k3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code             | kimi-k2.7-code             | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.6                  | kimi-k2.6                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Pro            | deepseek-v4-pro            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash          | deepseek-v4-flash          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5                  | mimo-v2.5                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro              | mimo-v2.5-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiniMax M3                 | minimax-m3                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.7               | minimax-m2.7               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.5               | minimax-m2.5               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Muse Spark 1.2 Contributor | muse-spark-1.2-contributor | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| Qwen3.8 Max                | qwen3.8-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Max                | qwen3.7-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus               | qwen3.7-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus               | qwen3.6-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Hy3                        | hy3                        | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Ox Alpha Free              | x-preview-f-free           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Model                        | Model ID                     | Endpoint                                         | AI SDK Package              |
+| ---------------------------- | ---------------------------- | ------------------------------------------------ | --------------------------- |
+| Grok 4.5                     | grok-4.5                     | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                 | gpt-5.6-luna                 | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GLM-5.3                      | glm-5.3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.2                      | glm-5.2                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.1                      | glm-5.1                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K3                      | kimi-k3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code               | kimi-k2.7-code               | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                    | kimi-k2.6                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Pro              | deepseek-v4-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash            | deepseek-v4-flash            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Vision Exp | deepseek-v4-flash-vision-exp | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5                    | mimo-v2.5                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro                | mimo-v2.5-pro                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiniMax M3                   | minimax-m3                   | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.7                 | minimax-m2.7                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.5                 | minimax-m2.5                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Muse Spark 1.2 Contributor   | muse-spark-1.2-contributor   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| Qwen3.8 Max                  | qwen3.8-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Max                  | qwen3.7-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                 | qwen3.7-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                 | qwen3.6-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Hy3                          | hy3                          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Ox Alpha Free                | ox-alpha-free                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |

 Dit [model id](/docs/config/#models) i din OpenCode config
 bruger formatet `opencode-go/<model-id>`. For eksempel for Kimi K3, vil du
@@ -254,29 +262,30 @@ https://opencode.ai/zen/go/v1/models

 ## Privatliv

-| Model                      | Modeltræning | Dataopbevaring |
-| -------------------------- | ------------ | -------------- |
-| Grok 4.5                   | Ikke brugt   | 30 dage        |
-| GPT 5.6 Luna               | Ikke brugt   | 30 dage        |
-| GLM-5.3                    | Ikke brugt   | 0 dage         |
-| GLM-5.2                    | Ikke brugt   | 0 dage         |
-| GLM-5.1                    | Ikke brugt   | 0 dage         |
-| Kimi K3                    | Ikke brugt   | 0 dage         |
-| Kimi K2.7 Code             | Ikke brugt   | 0 dage         |
-| Kimi K2.6                  | Ikke brugt   | 0 dage         |
-| MiMo-V2.5-Pro              | Ikke brugt   | 0 dage         |
-| MiMo-V2.5                  | Ikke brugt   | 0 dage         |
-| Qwen3.8 Max                | Ikke brugt   | 0 dage         |
-| Qwen3.7 Max                | Ikke brugt   | 0 dage         |
-| Qwen3.7 Plus               | Ikke brugt   | 0 dage         |
-| Qwen3.6 Plus               | Ikke brugt   | 0 dage         |
-| MiniMax M3                 | Ikke brugt   | 0 dage         |
-| MiniMax M2.7               | Ikke brugt   | 0 dage         |
-| Muse Spark 1.2 Contributor | Ja           | Ikke ZDR       |
-| DeepSeek V4 Pro            | Ikke brugt   | 0 dage         |
-| DeepSeek V4 Flash          | Ikke brugt   | 0 dage         |
-| Hy3                        | Ikke brugt   | 0 dage         |
-| Ox Alpha Free              | Ikke brugt   | 0 dage         |
+| Model                        | Modeltræning | Dataopbevaring |
+| ---------------------------- | ------------ | -------------- |
+| Grok 4.5                     | Ikke brugt   | 30 dage        |
+| GPT 5.6 Luna                 | Ikke brugt   | 30 dage        |
+| GLM-5.3                      | Ikke brugt   | 0 dage         |
+| GLM-5.2                      | Ikke brugt   | 0 dage         |
+| GLM-5.1                      | Ikke brugt   | 0 dage         |
+| Kimi K3                      | Ikke brugt   | 0 dage         |
+| Kimi K2.7 Code               | Ikke brugt   | 0 dage         |
+| Kimi K2.6                    | Ikke brugt   | 0 dage         |
+| MiMo-V2.5-Pro                | Ikke brugt   | 0 dage         |
+| MiMo-V2.5                    | Ikke brugt   | 0 dage         |
+| Qwen3.8 Max                  | Ikke brugt   | 0 dage         |
+| Qwen3.7 Max                  | Ikke brugt   | 0 dage         |
+| Qwen3.7 Plus                 | Ikke brugt   | 0 dage         |
+| Qwen3.6 Plus                 | Ikke brugt   | 0 dage         |
+| MiniMax M3                   | Ikke brugt   | 0 dage         |
+| MiniMax M2.7                 | Ikke brugt   | 0 dage         |
+| Muse Spark 1.2 Contributor   | Ja           | Ikke ZDR       |
+| DeepSeek V4 Pro              | Ikke brugt   | 0 dage         |
+| DeepSeek V4 Flash            | Ikke brugt   | 0 dage         |
+| DeepSeek V4 Flash Vision Exp | Ikke brugt   | 0 dage         |
+| Hy3                          | Ikke brugt   | 0 dage         |
+| Ox Alpha Free                | Ikke brugt   | 0 dage         |

 - **Grok 4.5:** ZDR deaktiverer vigtige API-funktioner, der afhænger af lagrede data, herunder den tilstandsbevarende Responses API, Files and Collections og Batch API. [Læs mere](https://docs.x.ai/developers/faq/security#what-is-zero-data-retention-zdr).
 - **GPT 5.6 Luna:** Logfiler til overvågning af misbrug genereres ved al brug af API-funktioner og opbevares i op til 30 dage. [Læs mere](https://developers.openai.com/api/docs/guides/your-data#data-retention-controls-for-abuse-monitoring).
@@ -195,8 +195,8 @@ Vi understøtter en pay-as-you-go-model. Nedenfor er priserne **pr. 1M tokens**.
 | Grok 4.5 (> 200K tokens)          | $4.00  | $12.00  | $0.60       | -            |
 | Grok Build 0.1                    | $1.00  | $2.00   | $0.20       | -            |
 | Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
-| GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
-| GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
+| GPT 5.6 Sol (≤ 272K tokens)       | $2.00  | $10.00  | $0.20       | $2.50        |
+| GPT 5.6 Sol (> 272K tokens)       | $4.00  | $15.00  | $0.40       | $5.00        |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
 | GPT 5.6 Terra (> 272K tokens)     | $4.00  | $18.00  | $0.40       | $5.00        |
 | GPT 5.6 Luna (≤ 272K tokens)      | $0.20  | $1.20   | $0.02       | $0.25        |
@@ -221,6 +221,8 @@ Vi understøtter en pay-as-you-go-model. Nedenfor er priserne **pr. 1M tokens**.
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |

+**GPT 5.6 Sol:** De viste priser inkluderer 50 % rabat til og med den 18. september 2026.
+
 **DeepSeek V4 Flash / Pro:** Peak-tiderne er 01:00-04:00 og 06:00-10:00 UTC; alle andre tider er Off-Peak. [Læs mere](https://api-docs.deepseek.com/quick_start/pricing/).

 Du vil måske bemærke [lavprismodeller](/docs/config/#models), såsom Haiku, Nano eller Flash, i din brugshistorik. OpenCode bruger disse modeller til at generere sessionstitler.
@@ -70,6 +70,7 @@ Die aktuelle Liste der Modelle umfasst:
 - **Qwen3.6 Plus**
 - **DeepSeek V4 Pro**
 - **DeepSeek V4 Flash**
+- **DeepSeek V4 Flash Vision Exp**
 - **Hy3**
 - **Ox Alpha Free** (für begrenzte Zeit)

@@ -89,29 +90,30 @@ Limits sind in Dollarwerten definiert. Das bedeutet, dass die tatsächliche Anza

 Die folgende Tabelle zeigt eine geschätzte Anzahl von Anfragen basierend auf typischen Go-Nutzungsmustern:

-| Model                      | Anfragen pro 5 Stunden | Anfragen pro Woche | Anfragen pro Monat |
-| -------------------------- | ---------------------- | ------------------ | ------------------ |
-| Grok 4.5                   | 120                    | 300                | 600                |
-| GPT 5.6 Luna               | 2,050                  | 5,100              | 10,250             |
-| GLM-5.3                    | 220                    | 540                | 1,080              |
-| GLM-5.2                    | 880                    | 2,150              | 4,300              |
-| GLM-5.1                    | 880                    | 2,150              | 4,300              |
-| Kimi K3                    | 110                    | 250                | 490                |
-| Kimi K2.7 Code             | 1,350                  | 3,380              | 6,750              |
-| Kimi K2.6                  | 1,150                  | 2,880              | 5,750              |
-| MiMo-V2.5                  | 30,100                 | 75,200             | 150,400            |
-| MiMo-V2.5-Pro              | 3,250                  | 8,150              | 16,300             |
-| MiniMax M3                 | 3,200                  | 8,000              | 16,000             |
-| MiniMax M2.7               | 3,400                  | 8,500              | 17,000             |
-| Muse Spark 1.2 Contributor | 45,300                 | 113,300            | 226,600            |
-| Qwen3.8 Max                | 160                    | 400                | 810                |
-| Qwen3.7 Max                | 340                    | 840                | 1,690              |
-| Qwen3.7 Plus               | 4,300                  | 10,800             | 21,600             |
-| Qwen3.6 Plus               | 3,300                  | 8,200              | 16,300             |
-| DeepSeek V4 Pro            | 1,050                  | 2,600              | 5,200              |
-| DeepSeek V4 Flash          | 7,600                  | 18,900             | 37,800             |
-| Hy3                        | 4,300                  | 10,750             | 21,500             |
-| Ox Alpha Free              | -                      | -                  | -                  |
+| Model                        | Anfragen pro 5 Stunden | Anfragen pro Woche | Anfragen pro Monat |
+| ---------------------------- | ---------------------- | ------------------ | ------------------ |
+| Grok 4.5                     | 120                    | 300                | 600                |
+| GPT 5.6 Luna                 | 2,050                  | 5,100              | 10,250             |
+| GLM-5.3                      | 220                    | 540                | 1,080              |
+| GLM-5.2                      | 880                    | 2,150              | 4,300              |
+| GLM-5.1                      | 880                    | 2,150              | 4,300              |
+| Kimi K3                      | 110                    | 250                | 490                |
+| Kimi K2.7 Code               | 1,350                  | 3,380              | 6,750              |
+| Kimi K2.6                    | 1,150                  | 2,880              | 5,750              |
+| MiMo-V2.5                    | 30,100                 | 75,200             | 150,400            |
+| MiMo-V2.5-Pro                | 3,250                  | 8,150              | 16,300             |
+| MiniMax M3                   | 3,200                  | 8,000              | 16,000             |
+| MiniMax M2.7                 | 3,400                  | 8,500              | 17,000             |
+| Muse Spark 1.2 Contributor   | 45,300                 | 113,300            | 226,600            |
+| Qwen3.8 Max                  | 160                    | 400                | 810                |
+| Qwen3.7 Max                  | 340                    | 840                | 1,690              |
+| Qwen3.7 Plus                 | 4,300                  | 10,800             | 21,600             |
+| Qwen3.6 Plus                 | 3,300                  | 8,200              | 16,300             |
+| DeepSeek V4 Pro              | 1,050                  | 2,600              | 5,200              |
+| DeepSeek V4 Flash            | 7,600                  | 18,900             | 37,800             |
+| DeepSeek V4 Flash Vision Exp | 3,800                  | 9,450              | 18,900             |
+| Hy3                          | 4,300                  | 10,750             | 21,500             |
+| Ox Alpha Free                | -                      | -                  | -                  |

 Die Schätzungen basieren auf beobachteten Anfragemustern:

@@ -122,6 +124,7 @@ Die Schätzungen basieren auf beobachteten Anfragemustern:
 - Kimi K2.7/K2.6 — 870 Input-, 55.000 Cached-, 200 Output-Tokens pro Anfrage
 - DeepSeek V4 Pro — 750 Input-, 82.000 Cached-, 290 Output-Tokens pro Anfrage
 - DeepSeek V4 Flash — 410 Input-, 71.300 Cached-, 310 Output-Tokens pro Anfrage
+- DeepSeek V4 Flash Vision Exp — 410 Input-, 71.300 Cached-, 310 Output-Tokens pro Anfrage
 - MiniMax M3 — 510 Input-, 56.000 Cached-, 190 Output-Tokens pro Anfrage
 - MiniMax M2.7 — 300 Input-, 55.000 Cached-, 125 Output-Tokens pro Anfrage
 - Muse Spark 1.2 Contributor — 620 Input-, 71.400 Cached-, 300 Output-Tokens pro Anfrage
@@ -135,37 +138,41 @@ Die Schätzungen basieren auf beobachteten Anfragemustern:

 Die Schätzungen basieren außerdem auf den folgenden Preisen pro 1M Tokens und der monatlichen Nutzung, die bei jedem Modell enthalten ist:

-| Model                        | Input  | Output | Cached Read | Cached Write | Nutzung |
-| ---------------------------- | ------ | ------ | ----------- | ------------ | ------- |
-| Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15     |
-| GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15     |
-| GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15     |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15     |
-| GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60     |
-| GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60     |
-| Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15     |
-| Kimi K2.7 Code               | $0.95  | $4.00  | $0.19       | -            | $60     |
-| Kimi K2.6                    | $0.95  | $4.00  | $0.16       | -            | $60     |
-| MiMo V2.5                    | $0.14  | $0.28  | $0.0028     | -            | $60     |
-| MiMo V2.5 Pro                | $0.435 | $0.87  | $0.003625   | -            | $15     |
-| MiniMax M3                   | $0.30  | $1.20  | $0.06       | -            | $60     |
-| MiniMax M2.7                 | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
-| MiniMax M2.5                 | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
-| Muse Spark 1.2 Contributor   | $0.10  | $0.20  | $0.002      | -            | $60     |
-| Qwen3.8 Max                  | $2.00  | $6.00  | $0.25       | $2.50        | $15     |
-| Qwen3.7 Max                  | $2.50  | $7.50  | $0.50       | $3.125       | $60     |
-| Qwen3.7 Plus (≤ 256K tokens) | $0.40  | $1.60  | $0.04       | $0.50        | $60     |
-| Qwen3.7 Plus (> 256K tokens) | $1.20  | $4.80  | $0.12       | $1.50        | $60     |
-| Qwen3.6 Plus (≤ 256K tokens) | $0.50  | $3.00  | $0.05       | $0.625       | $60     |
-| Qwen3.6 Plus (> 256K tokens) | $2.00  | $6.00  | $0.20       | $2.50        | $60     |
-| DeepSeek V4 Pro (Off-Peak)   | $0.66  | $1.98  | $0.022      | -            | $15     |
-| DeepSeek V4 Pro (Peak)       | $1.32  | $3.96  | $0.044      | -            | $15     |
-| DeepSeek V4 Flash (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $30     |
-| DeepSeek V4 Flash (Peak)     | $0.44  | $1.32  | $0.014      | -            | $30     |
-| Hy3                          | $0.14  | $0.58  | $0.035      | -            | $60     |
-| Ox Alpha Free                | -      | -      | -           | -            | -       |
+| Model                                   | Input  | Output | Cached Read | Cached Write | Nutzung |
+| --------------------------------------- | ------ | ------ | ----------- | ------------ | ------- |
+| Grok 4.5                                | $2.00  | $6.00  | $0.30       | -            | $15     |
+| GPT 5.6 Luna (≤ 272K tokens)            | $0.20  | $1.20  | $0.02       | $0.25        | $15     |
+| GPT 5.6 Luna (> 272K tokens)            | $0.40  | $1.80  | $0.04       | $0.50        | $15     |
+| GLM-5.3                                 | $1.40  | $4.40  | $0.26       | -            | $15     |
+| GLM-5.2                                 | $1.40  | $4.40  | $0.26       | -            | $60     |
+| GLM-5.1                                 | $1.40  | $4.40  | $0.26       | -            | $60     |
+| Kimi K3                                 | $3.00  | $15.00 | $0.30       | -            | $15     |
+| Kimi K2.7 Code                          | $0.95  | $4.00  | $0.19       | -            | $60     |
+| Kimi K2.6                               | $0.95  | $4.00  | $0.16       | -            | $60     |
+| MiMo V2.5                               | $0.14  | $0.28  | $0.0028     | -            | $60     |
+| MiMo V2.5 Pro                           | $0.435 | $0.87  | $0.003625   | -            | $15     |
+| MiniMax M3                              | $0.30  | $1.20  | $0.06       | -            | $60     |
+| MiniMax M2.7                            | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
+| MiniMax M2.5                            | $0.30  | $1.20  | $0.06       | $0.375       | $60     |
+| Muse Spark 1.2 Contributor              | $0.10  | $0.20  | $0.002      | -            | $60     |
+| Qwen3.8 Max                             | $2.00  | $6.00  | $0.25       | $2.50        | $15     |
+| Qwen3.7 Max                             | $2.50  | $7.50  | $0.50       | $3.125       | $60     |
+| Qwen3.7 Plus (≤ 256K tokens)            | $0.40  | $1.60  | $0.04       | $0.50        | $60     |
+| Qwen3.7 Plus (> 256K tokens)            | $1.20  | $4.80  | $0.12       | $1.50        | $60     |
+| Qwen3.6 Plus (≤ 256K tokens)            | $0.50  | $3.00  | $0.05       | $0.625       | $60     |
+| Qwen3.6 Plus (> 256K tokens)            | $2.00  | $6.00  | $0.20       | $2.50        | $60     |
+| DeepSeek V4 Pro (Off-Peak)              | $0.66  | $1.98  | $0.022      | -            | $15     |
+| DeepSeek V4 Pro (Peak)                  | $1.32  | $3.96  | $0.044      | -            | $15     |
+| DeepSeek V4 Flash (Off-Peak)            | $0.22  | $0.66  | $0.007      | -            | $30     |
+| DeepSeek V4 Flash (Peak)                | $0.44  | $1.32  | $0.014      | -            | $30     |
+| DeepSeek V4 Flash Vision Exp (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $15     |
+| DeepSeek V4 Flash Vision Exp (Peak)     | $0.44  | $1.32  | $0.014      | -            | $15     |
+| Hy3                                     | $0.14  | $0.58  | $0.035      | -            | $60     |
+| Ox Alpha Free                           | -      | -      | -           | -            | -       |

-**DeepSeek V4 Flash / Pro:** Die Peak-Zeiten sind 01:00-04:00 und 06:00-10:00 UTC; alle anderen Zeiten sind Off-Peak. [Mehr erfahren](https://api-docs.deepseek.com/quick_start/pricing/).
+**DeepSeek V4 Flash / V4 Flash Vision Exp / V4 Pro:** Die Peak-Zeiten sind 01:00-04:00 und 06:00-10:00 UTC; alle anderen Zeiten sind Off-Peak. [Mehr erfahren](https://api-docs.deepseek.com/quick_start/pricing/).
+
+**DeepSeek V4 Flash Vision Exp:** Bilder werden anhand ihrer Abmessungen in Tokens umgewandelt und zusammen mit Text-Tokens als Input-Tokens abgerechnet. [Mehr erfahren](https://api-docs.deepseek.com/quick_start/pricing/).

 **Ox Alpha Free:** Für begrenzte Zeit kostenlos.

@@ -201,30 +208,31 @@ Bei diesen Modellen erhältst du immer noch etwas mehr, als wenn du die Modellan

 Du kannst auf die Go-Modelle auch über die folgenden API-Endpunkte zugreifen.

-| Modell                     | Modell-ID                  | Endpunkt                                         | AI SDK Package              |
-| -------------------------- | -------------------------- | ------------------------------------------------ | --------------------------- |
-| Grok 4.5                   | grok-4.5                   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GPT 5.6 Luna               | gpt-5.6-luna               | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3                    | glm-5.3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.2                    | glm-5.2                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.1                    | glm-5.1                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K3                    | kimi-k3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code             | kimi-k2.7-code             | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.6                  | kimi-k2.6                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Pro            | deepseek-v4-pro            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash          | deepseek-v4-flash          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5                  | mimo-v2.5                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro              | mimo-v2.5-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiniMax M3                 | minimax-m3                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.7               | minimax-m2.7               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.5               | minimax-m2.5               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Muse Spark 1.2 Contributor | muse-spark-1.2-contributor | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| Qwen3.8 Max                | qwen3.8-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Max                | qwen3.7-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus               | qwen3.7-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus               | qwen3.6-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Hy3                        | hy3                        | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Ox Alpha Free              | x-preview-f-free           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Modell                       | Modell-ID                    | Endpunkt                                         | AI SDK Package              |
+| ---------------------------- | ---------------------------- | ------------------------------------------------ | --------------------------- |
+| Grok 4.5                     | grok-4.5                     | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                 | gpt-5.6-luna                 | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GLM-5.3                      | glm-5.3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.2                      | glm-5.2                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.1                      | glm-5.1                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K3                      | kimi-k3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code               | kimi-k2.7-code               | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                    | kimi-k2.6                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Pro              | deepseek-v4-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash            | deepseek-v4-flash            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Vision Exp | deepseek-v4-flash-vision-exp | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5                    | mimo-v2.5                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro                | mimo-v2.5-pro                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiniMax M3                   | minimax-m3                   | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.7                 | minimax-m2.7                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.5                 | minimax-m2.5                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Muse Spark 1.2 Contributor   | muse-spark-1.2-contributor   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| Qwen3.8 Max                  | qwen3.8-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Max                  | qwen3.7-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                 | qwen3.7-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                 | qwen3.6-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Hy3                          | hy3                          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Ox Alpha Free                | ox-alpha-free                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |

 Die [Modell-ID](/docs/config/#models) in deiner OpenCode Config verwendet das Format `opencode-go/<model-id>`. Für Kimi K3 würdest du beispielsweise `opencode-go/kimi-k3` in deiner Config verwenden.

@@ -242,29 +250,30 @@ https://opencode.ai/zen/go/v1/models

 ## Datenschutz

-| Modell                     | Modelltraining  | Datenaufbewahrung |
-| -------------------------- | --------------- | ----------------- |
-| Grok 4.5                   | Nicht verwendet | 30 Tage           |
-| GPT 5.6 Luna               | Nicht verwendet | 30 Tage           |
-| GLM-5.3                    | Nicht verwendet | 0 Tage            |
-| GLM-5.2                    | Nicht verwendet | 0 Tage            |
-| GLM-5.1                    | Nicht verwendet | 0 Tage            |
-| Kimi K3                    | Nicht verwendet | 0 Tage            |
-| Kimi K2.7 Code             | Nicht verwendet | 0 Tage            |
-| Kimi K2.6                  | Nicht verwendet | 0 Tage            |
-| MiMo-V2.5-Pro              | Nicht verwendet | 0 Tage            |
-| MiMo-V2.5                  | Nicht verwendet | 0 Tage            |
-| Qwen3.8 Max                | Nicht verwendet | 0 Tage            |
-| Qwen3.7 Max                | Nicht verwendet | 0 Tage            |
-| Qwen3.7 Plus               | Nicht verwendet | 0 Tage            |
-| Qwen3.6 Plus               | Nicht verwendet | 0 Tage            |
-| MiniMax M3                 | Nicht verwendet | 0 Tage            |
-| MiniMax M2.7               | Nicht verwendet | 0 Tage            |
-| Muse Spark 1.2 Contributor | Ja              | Kein ZDR          |
-| DeepSeek V4 Pro            | Nicht verwendet | 0 Tage            |
-| DeepSeek V4 Flash          | Nicht verwendet | 0 Tage            |
-| Hy3                        | Nicht verwendet | 0 Tage            |
-| Ox Alpha Free              | Nicht verwendet | 0 Tage            |
+| Modell                       | Modelltraining  | Datenaufbewahrung |
+| ---------------------------- | --------------- | ----------------- |
+| Grok 4.5                     | Nicht verwendet | 30 Tage           |
+| GPT 5.6 Luna                 | Nicht verwendet | 30 Tage           |
+| GLM-5.3                      | Nicht verwendet | 0 Tage            |
+| GLM-5.2                      | Nicht verwendet | 0 Tage            |
+| GLM-5.1                      | Nicht verwendet | 0 Tage            |
+| Kimi K3                      | Nicht verwendet | 0 Tage            |
+| Kimi K2.7 Code               | Nicht verwendet | 0 Tage            |
+| Kimi K2.6                    | Nicht verwendet | 0 Tage            |
+| MiMo-V2.5-Pro                | Nicht verwendet | 0 Tage            |
+| MiMo-V2.5                    | Nicht verwendet | 0 Tage            |
+| Qwen3.8 Max                  | Nicht verwendet | 0 Tage            |
+| Qwen3.7 Max                  | Nicht verwendet | 0 Tage            |
+| Qwen3.7 Plus                 | Nicht verwendet | 0 Tage            |
+| Qwen3.6 Plus                 | Nicht verwendet | 0 Tage            |
+| MiniMax M3                   | Nicht verwendet | 0 Tage            |
+| MiniMax M2.7                 | Nicht verwendet | 0 Tage            |
+| Muse Spark 1.2 Contributor   | Ja              | Kein ZDR          |
+| DeepSeek V4 Pro              | Nicht verwendet | 0 Tage            |
+| DeepSeek V4 Flash            | Nicht verwendet | 0 Tage            |
+| DeepSeek V4 Flash Vision Exp | Nicht verwendet | 0 Tage            |
+| Hy3                          | Nicht verwendet | 0 Tage            |
+| Ox Alpha Free                | Nicht verwendet | 0 Tage            |

 - **Grok 4.5:** ZDR deaktiviert wichtige API-Funktionen, die von gespeicherten Daten abhängen, einschließlich der zustandsbehafteten Responses API, Files and Collections und der Batch API. [Mehr erfahren](https://docs.x.ai/developers/faq/security#what-is-zero-data-retention-zdr).
 - **GPT 5.6 Luna:** Für die Nutzung aller API-Funktionen werden Protokolle zur Missbrauchsüberwachung erstellt und bis zu 30 Tage lang aufbewahrt. [Mehr erfahren](https://developers.openai.com/api/docs/guides/your-data#data-retention-controls-for-abuse-monitoring).
@@ -184,8 +184,8 @@ Wir unterstützen ein Pay-as-you-go-Modell. Unten findest du die Preise **pro 1M
 | Grok 4.5 (> 200K tokens)          | $4.00  | $12.00  | $0.60       | -            |
 | Grok Build 0.1                    | $1.00  | $2.00   | $0.20       | -            |
 | Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
-| GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
-| GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
+| GPT 5.6 Sol (≤ 272K tokens)       | $2.00  | $10.00  | $0.20       | $2.50        |
+| GPT 5.6 Sol (> 272K tokens)       | $4.00  | $15.00  | $0.40       | $5.00        |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
 | GPT 5.6 Terra (> 272K tokens)     | $4.00  | $18.00  | $0.40       | $5.00        |
 | GPT 5.6 Luna (≤ 272K tokens)      | $0.20  | $1.20   | $0.02       | $0.25        |
@@ -210,6 +210,8 @@ Wir unterstützen ein Pay-as-you-go-Modell. Unten findest du die Preise **pro 1M
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |

+**GPT 5.6 Sol:** Die angezeigten Preise enthalten bis zum 18. September 2026 einen Rabatt von 50 %.
+
 **DeepSeek V4 Flash / Pro:** Die Peak-Zeiten sind 01:00-04:00 und 06:00-10:00 UTC; alle anderen Zeiten sind Off-Peak. [Mehr erfahren](https://api-docs.deepseek.com/quick_start/pricing/).

 Möglicherweise siehst du [kostengünstige Modelle](/docs/config/#models) wie Haiku, Nano oder Flash in deinem Nutzungsverlauf. OpenCode verwendet diese Modelle, um Session-Titel zu generieren.
@@ -78,6 +78,7 @@ La lista actual de modelos incluye:
 - **Qwen3.6 Plus**
 - **DeepSeek V4 Pro**
 - **DeepSeek V4 Flash**
+- **DeepSeek V4 Flash Vision Exp**
 - **Hy3**
 - **Ox Alpha Free** (por tiempo limitado)

@@ -97,29 +98,30 @@ Los límites se definen en valor en dólares. Esto significa que tu cantidad rea

 La siguiente tabla proporciona una cantidad estimada de peticiones basada en los patrones típicos de uso de Go:

-| Model                      | peticiones por 5 horas | peticiones por semana | peticiones por mes |
-| -------------------------- | ---------------------- | --------------------- | ------------------ |
-| Grok 4.5                   | 120                    | 300                   | 600                |
-| GPT 5.6 Luna               | 2,050                  | 5,100                 | 10,250             |
-| GLM-5.3                    | 220                    | 540                   | 1,080              |
-| GLM-5.2                    | 880                    | 2,150                 | 4,300              |
-| GLM-5.1                    | 880                    | 2,150                 | 4,300              |
-| Kimi K3                    | 110                    | 250                   | 490                |
-| Kimi K2.7 Code             | 1,350                  | 3,380                 | 6,750              |
-| Kimi K2.6                  | 1,150                  | 2,880                 | 5,750              |
-| MiMo-V2.5                  | 30,100                 | 75,200                | 150,400            |
-| MiMo-V2.5-Pro              | 3,250                  | 8,150                 | 16,300             |
-| MiniMax M3                 | 3,200                  | 8,000                 | 16,000             |
-| MiniMax M2.7               | 3,400                  | 8,500                 | 17,000             |
-| Muse Spark 1.2 Contributor | 45,300                 | 113,300               | 226,600            |
-| Qwen3.8 Max                | 160                    | 400                   | 810                |
-| Qwen3.7 Max                | 340                    | 840                   | 1,690              |
-| Qwen3.7 Plus               | 4,300                  | 10,800                | 21,600             |
-| Qwen3.6 Plus               | 3,300                  | 8,200                 | 16,300             |
-| DeepSeek V4 Pro            | 1,050                  | 2,600                 | 5,200              |
-| DeepSeek V4 Flash          | 7,600                  | 18,900                | 37,800             |
-| Hy3                        | 4,300                  | 10,750                | 21,500             |
-| Ox Alpha Free              | -                      | -                     | -                  |
+| Model                        | peticiones por 5 horas | peticiones por semana | peticiones por mes |
+| ---------------------------- | ---------------------- | --------------------- | ------------------ |
+| Grok 4.5                     | 120                    | 300                   | 600                |
+| GPT 5.6 Luna                 | 2,050                  | 5,100                 | 10,250             |
+| GLM-5.3                      | 220                    | 540                   | 1,080              |
+| GLM-5.2                      | 880                    | 2,150                 | 4,300              |
+| GLM-5.1                      | 880                    | 2,150                 | 4,300              |
+| Kimi K3                      | 110                    | 250                   | 490                |
+| Kimi K2.7 Code               | 1,350                  | 3,380                 | 6,750              |
+| Kimi K2.6                    | 1,150                  | 2,880                 | 5,750              |
+| MiMo-V2.5                    | 30,100                 | 75,200                | 150,400            |
+| MiMo-V2.5-Pro                | 3,250                  | 8,150                 | 16,300             |
+| MiniMax M3                   | 3,200                  | 8,000                 | 16,000             |
+| MiniMax M2.7                 | 3,400                  | 8,500                 | 17,000             |
+| Muse Spark 1.2 Contributor   | 45,300                 | 113,300               | 226,600            |
+| Qwen3.8 Max                  | 160                    | 400                   | 810                |
+| Qwen3.7 Max                  | 340                    | 840                   | 1,690              |
+| Qwen3.7 Plus                 | 4,300                  | 10,800                | 21,600             |
+| Qwen3.6 Plus                 | 3,300                  | 8,200                 | 16,300             |
+| DeepSeek V4 Pro              | 1,050                  | 2,600                 | 5,200              |
+| DeepSeek V4 Flash            | 7,600                  | 18,900                | 37,800             |
+| DeepSeek V4 Flash Vision Exp | 3,800                  | 9,450                 | 18,900             |
+| Hy3                          | 4,300                  | 10,750                | 21,500             |
+| Ox Alpha Free                | -                      | -                     | -                  |

 Las estimaciones se basan en los patrones de peticiones observados:

@@ -130,6 +132,7 @@ Las estimaciones se basan en los patrones de peticiones observados:
 - Kimi K2.7/K2.6 — 870 tokens de entrada, 55,000 en caché, 200 tokens de salida por petición
 - DeepSeek V4 Pro — 750 tokens de entrada, 82,000 en caché, 290 tokens de salida por petición
 - DeepSeek V4 Flash — 410 tokens de entrada, 71,300 en caché, 310 tokens de salida por petición
+- DeepSeek V4 Flash Vision Exp — 410 tokens de entrada, 71,300 en caché, 310 tokens de salida por petición
 - MiniMax M3 — 510 tokens de entrada, 56,000 en caché, 190 tokens de salida por petición
 - MiniMax M2.7 — 300 tokens de entrada, 55,000 en caché, 125 tokens de salida por petición
 - Muse Spark 1.2 Contributor — 620 tokens de entrada, 71,400 en caché, 300 tokens de salida por petición
@@ -143,37 +146,41 @@ Las estimaciones se basan en los patrones de peticiones observados:

 Las estimaciones también se basan en los siguientes precios por 1M tokens y en el uso mensual incluido con cada modelo:

-| Modelo                       | Entrada | Salida | Lectura en caché | Escritura en caché | Uso |
-| ---------------------------- | ------- | ------ | ---------------- | ------------------ | --- |
-| Grok 4.5                     | $2.00   | $6.00  | $0.30            | -                  | $15 |
-| GPT 5.6 Luna (≤ 272K tokens) | $0.20   | $1.20  | $0.02            | $0.25              | $15 |
-| GPT 5.6 Luna (> 272K tokens) | $0.40   | $1.80  | $0.04            | $0.50              | $15 |
-| GLM-5.3                      | $1.40   | $4.40  | $0.26            | -                  | $15 |
-| GLM-5.2                      | $1.40   | $4.40  | $0.26            | -                  | $60 |
-| GLM-5.1                      | $1.40   | $4.40  | $0.26            | -                  | $60 |
-| Kimi K3                      | $3.00   | $15.00 | $0.30            | -                  | $15 |
-| Kimi K2.7 Code               | $0.95   | $4.00  | $0.19            | -                  | $60 |
-| Kimi K2.6                    | $0.95   | $4.00  | $0.16            | -                  | $60 |
-| MiMo V2.5                    | $0.14   | $0.28  | $0.0028          | -                  | $60 |
-| MiMo V2.5 Pro                | $0.435  | $0.87  | $0.003625        | -                  | $15 |
-| MiniMax M3                   | $0.30   | $1.20  | $0.06            | -                  | $60 |
-| MiniMax M2.7                 | $0.30   | $1.20  | $0.06            | $0.375             | $60 |
-| MiniMax M2.5                 | $0.30   | $1.20  | $0.06            | $0.375             | $60 |
-| Muse Spark 1.2 Contributor   | $0.10   | $0.20  | $0.002           | -                  | $60 |
-| Qwen3.8 Max                  | $2.00   | $6.00  | $0.25            | $2.50              | $15 |
-| Qwen3.7 Max                  | $2.50   | $7.50  | $0.50            | $3.125             | $60 |
-| Qwen3.7 Plus (≤ 256K tokens) | $0.40   | $1.60  | $0.04            | $0.50              | $60 |
-| Qwen3.7 Plus (> 256K tokens) | $1.20   | $4.80  | $0.12            | $1.50              | $60 |
-| Qwen3.6 Plus (≤ 256K tokens) | $0.50   | $3.00  | $0.05            | $0.625             | $60 |
-| Qwen3.6 Plus (> 256K tokens) | $2.00   | $6.00  | $0.20            | $2.50              | $60 |
-| DeepSeek V4 Pro (Off-Peak)   | $0.66   | $1.98  | $0.022           | -                  | $15 |
-| DeepSeek V4 Pro (Peak)       | $1.32   | $3.96  | $0.044           | -                  | $15 |
-| DeepSeek V4 Flash (Off-Peak) | $0.22   | $0.66  | $0.007           | -                  | $30 |
-| DeepSeek V4 Flash (Peak)     | $0.44   | $1.32  | $0.014           | -                  | $30 |
-| Hy3                          | $0.14   | $0.58  | $0.035           | -                  | $60 |
-| Ox Alpha Free                | -       | -      | -                | -                  | -   |
+| Modelo                                  | Entrada | Salida | Lectura en caché | Escritura en caché | Uso |
+| --------------------------------------- | ------- | ------ | ---------------- | ------------------ | --- |
+| Grok 4.5                                | $2.00   | $6.00  | $0.30            | -                  | $15 |
+| GPT 5.6 Luna (≤ 272K tokens)            | $0.20   | $1.20  | $0.02            | $0.25              | $15 |
+| GPT 5.6 Luna (> 272K tokens)            | $0.40   | $1.80  | $0.04            | $0.50              | $15 |
+| GLM-5.3                                 | $1.40   | $4.40  | $0.26            | -                  | $15 |
+| GLM-5.2                                 | $1.40   | $4.40  | $0.26            | -                  | $60 |
+| GLM-5.1                                 | $1.40   | $4.40  | $0.26            | -                  | $60 |
+| Kimi K3                                 | $3.00   | $15.00 | $0.30            | -                  | $15 |
+| Kimi K2.7 Code                          | $0.95   | $4.00  | $0.19            | -                  | $60 |
+| Kimi K2.6                               | $0.95   | $4.00  | $0.16            | -                  | $60 |
+| MiMo V2.5                               | $0.14   | $0.28  | $0.0028          | -                  | $60 |
+| MiMo V2.5 Pro                           | $0.435  | $0.87  | $0.003625        | -                  | $15 |
+| MiniMax M3                              | $0.30   | $1.20  | $0.06            | -                  | $60 |
+| MiniMax M2.7                            | $0.30   | $1.20  | $0.06            | $0.375             | $60 |
+| MiniMax M2.5                            | $0.30   | $1.20  | $0.06            | $0.375             | $60 |
+| Muse Spark 1.2 Contributor              | $0.10   | $0.20  | $0.002           | -                  | $60 |
+| Qwen3.8 Max                             | $2.00   | $6.00  | $0.25            | $2.50              | $15 |
+| Qwen3.7 Max                             | $2.50   | $7.50  | $0.50            | $3.125             | $60 |
+| Qwen3.7 Plus (≤ 256K tokens)            | $0.40   | $1.60  | $0.04            | $0.50              | $60 |
+| Qwen3.7 Plus (> 256K tokens)            | $1.20   | $4.80  | $0.12            | $1.50              | $60 |
+| Qwen3.6 Plus (≤ 256K tokens)            | $0.50   | $3.00  | $0.05            | $0.625             | $60 |
+| Qwen3.6 Plus (> 256K tokens)            | $2.00   | $6.00  | $0.20            | $2.50              | $60 |
+| DeepSeek V4 Pro (Off-Peak)              | $0.66   | $1.98  | $0.022           | -                  | $15 |
+| DeepSeek V4 Pro (Peak)                  | $1.32   | $3.96  | $0.044           | -                  | $15 |
+| DeepSeek V4 Flash (Off-Peak)            | $0.22   | $0.66  | $0.007           | -                  | $30 |
+| DeepSeek V4 Flash (Peak)                | $0.44   | $1.32  | $0.014           | -                  | $30 |
+| DeepSeek V4 Flash Vision Exp (Off-Peak) | $0.22   | $0.66  | $0.007           | -                  | $15 |
+| DeepSeek V4 Flash Vision Exp (Peak)     | $0.44   | $1.32  | $0.014           | -                  | $15 |
+| Hy3                                     | $0.14   | $0.58  | $0.035           | -                  | $60 |
+| Ox Alpha Free                           | -       | -      | -                | -                  | -   |

-**DeepSeek V4 Flash / Pro:** Las horas Peak son 01:00-04:00 y 06:00-10:00 UTC; todas las demás horas son Off-Peak. [Más información](https://api-docs.deepseek.com/quick_start/pricing/).
+**DeepSeek V4 Flash / V4 Flash Vision Exp / V4 Pro:** Las horas Peak son 01:00-04:00 y 06:00-10:00 UTC; todas las demás horas son Off-Peak. [Más información](https://api-docs.deepseek.com/quick_start/pricing/).
+
+**DeepSeek V4 Flash Vision Exp:** Las imágenes se convierten en tokens según sus dimensiones y se facturan como tokens de entrada junto con los tokens de texto. [Más información](https://api-docs.deepseek.com/quick_start/pricing/).

 **Ox Alpha Free:** Gratis por tiempo limitado.

@@ -211,30 +218,31 @@ Con estos modelos, aun así obtienes un poco más que si pagaras directamente a

 También puedes acceder a los modelos de Go a través de los siguientes endpoints de la API.

-| Modelo                     | ID del modelo              | Endpoint                                         | Paquete de AI SDK           |
-| -------------------------- | -------------------------- | ------------------------------------------------ | --------------------------- |
-| Grok 4.5                   | grok-4.5                   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GPT 5.6 Luna               | gpt-5.6-luna               | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3                    | glm-5.3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.2                    | glm-5.2                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.1                    | glm-5.1                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K3                    | kimi-k3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code             | kimi-k2.7-code             | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.6                  | kimi-k2.6                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Pro            | deepseek-v4-pro            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash          | deepseek-v4-flash          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5                  | mimo-v2.5                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro              | mimo-v2.5-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiniMax M3                 | minimax-m3                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.7               | minimax-m2.7               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.5               | minimax-m2.5               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Muse Spark 1.2 Contributor | muse-spark-1.2-contributor | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| Qwen3.8 Max                | qwen3.8-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Max                | qwen3.7-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus               | qwen3.7-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus               | qwen3.6-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Hy3                        | hy3                        | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Ox Alpha Free              | x-preview-f-free           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Modelo                       | ID del modelo                | Endpoint                                         | Paquete de AI SDK           |
+| ---------------------------- | ---------------------------- | ------------------------------------------------ | --------------------------- |
+| Grok 4.5                     | grok-4.5                     | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                 | gpt-5.6-luna                 | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GLM-5.3                      | glm-5.3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.2                      | glm-5.2                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.1                      | glm-5.1                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K3                      | kimi-k3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code               | kimi-k2.7-code               | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                    | kimi-k2.6                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Pro              | deepseek-v4-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash            | deepseek-v4-flash            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Vision Exp | deepseek-v4-flash-vision-exp | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5                    | mimo-v2.5                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro                | mimo-v2.5-pro                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiniMax M3                   | minimax-m3                   | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.7                 | minimax-m2.7                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.5                 | minimax-m2.5                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Muse Spark 1.2 Contributor   | muse-spark-1.2-contributor   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| Qwen3.8 Max                  | qwen3.8-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Max                  | qwen3.7-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                 | qwen3.7-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                 | qwen3.6-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Hy3                          | hy3                          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Ox Alpha Free                | ox-alpha-free                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |

 El [ID del modelo](/docs/config/#models) en tu configuración de OpenCode
 usa el formato `opencode-go/<model-id>`. Por ejemplo, para Kimi K3, usarías
@@ -254,29 +262,30 @@ https://opencode.ai/zen/go/v1/models

 ## Privacidad

-| Modelo                     | Entrenamiento del modelo | Retención de datos |
-| -------------------------- | ------------------------ | ------------------ |
-| Grok 4.5                   | No utilizado             | 30 días            |
-| GPT 5.6 Luna               | No utilizado             | 30 días            |
-| GLM-5.3                    | No utilizado             | 0 días             |
-| GLM-5.2                    | No utilizado             | 0 días             |
-| GLM-5.1                    | No utilizado             | 0 días             |
-| Kimi K3                    | No utilizado             | 0 días             |
-| Kimi K2.7 Code             | No utilizado             | 0 días             |
-| Kimi K2.6                  | No utilizado             | 0 días             |
-| MiMo-V2.5-Pro              | No utilizado             | 0 días             |
-| MiMo-V2.5                  | No utilizado             | 0 días             |
-| Qwen3.8 Max                | No utilizado             | 0 días             |
-| Qwen3.7 Max                | No utilizado             | 0 días             |
-| Qwen3.7 Plus               | No utilizado             | 0 días             |
-| Qwen3.6 Plus               | No utilizado             | 0 días             |
-| MiniMax M3                 | No utilizado             | 0 días             |
-| MiniMax M2.7               | No utilizado             | 0 días             |
-| Muse Spark 1.2 Contributor | Sí                       | Sin ZDR            |
-| DeepSeek V4 Pro            | No utilizado             | 0 días             |
-| DeepSeek V4 Flash          | No utilizado             | 0 días             |
-| Hy3                        | No utilizado             | 0 días             |
-| Ox Alpha Free              | No utilizado             | 0 días             |
+| Modelo                       | Entrenamiento del modelo | Retención de datos |
+| ---------------------------- | ------------------------ | ------------------ |
+| Grok 4.5                     | No utilizado             | 30 días            |
+| GPT 5.6 Luna                 | No utilizado             | 30 días            |
+| GLM-5.3                      | No utilizado             | 0 días             |
+| GLM-5.2                      | No utilizado             | 0 días             |
+| GLM-5.1                      | No utilizado             | 0 días             |
+| Kimi K3                      | No utilizado             | 0 días             |
+| Kimi K2.7 Code               | No utilizado             | 0 días             |
+| Kimi K2.6                    | No utilizado             | 0 días             |
+| MiMo-V2.5-Pro                | No utilizado             | 0 días             |
+| MiMo-V2.5                    | No utilizado             | 0 días             |
+| Qwen3.8 Max                  | No utilizado             | 0 días             |
+| Qwen3.7 Max                  | No utilizado             | 0 días             |
+| Qwen3.7 Plus                 | No utilizado             | 0 días             |
+| Qwen3.6 Plus                 | No utilizado             | 0 días             |
+| MiniMax M3                   | No utilizado             | 0 días             |
+| MiniMax M2.7                 | No utilizado             | 0 días             |
+| Muse Spark 1.2 Contributor   | Sí                       | Sin ZDR            |
+| DeepSeek V4 Pro              | No utilizado             | 0 días             |
+| DeepSeek V4 Flash            | No utilizado             | 0 días             |
+| DeepSeek V4 Flash Vision Exp | No utilizado             | 0 días             |
+| Hy3                          | No utilizado             | 0 días             |
+| Ox Alpha Free                | No utilizado             | 0 días             |

 - **Grok 4.5:** ZDR deshabilita funciones importantes de la API que dependen de datos almacenados, incluidas la Responses API con estado, Files and Collections y la Batch API. [Más información](https://docs.x.ai/developers/faq/security#what-is-zero-data-retention-zdr).
 - **GPT 5.6 Luna:** Se generan registros de supervisión de abusos para todo el uso de funciones de la API y se conservan durante un máximo de 30 días. [Más información](https://developers.openai.com/api/docs/guides/your-data#data-retention-controls-for-abuse-monitoring).
@@ -195,8 +195,8 @@ Admitimos un modelo de pago por uso. A continuación se muestran los precios **p
 | Grok 4.5 (> 200K tokens)          | $4.00   | $12.00  | $0.60            | -                  |
 | Grok Build 0.1                    | $1.00   | $2.00   | $0.20            | -                  |
 | Muse Spark 1.2                    | $1.25   | $4.25   | $0.15            | -                  |
-| GPT 5.6 Sol (≤ 272K tokens)       | $5.00   | $30.00  | $0.50            | $6.25              |
-| GPT 5.6 Sol (> 272K tokens)       | $10.00  | $45.00  | $1.00            | $12.50             |
+| GPT 5.6 Sol (≤ 272K tokens)       | $2.00   | $10.00  | $0.20            | $2.50              |
+| GPT 5.6 Sol (> 272K tokens)       | $4.00   | $15.00  | $0.40            | $5.00              |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00   | $12.00  | $0.20            | $2.50              |
 | GPT 5.6 Terra (> 272K tokens)     | $4.00   | $18.00  | $0.40            | $5.00              |
 | GPT 5.6 Luna (≤ 272K tokens)      | $0.20   | $1.20   | $0.02            | $0.25              |
@@ -221,6 +221,8 @@ Admitimos un modelo de pago por uso. A continuación se muestran los precios **p
 | GPT 5 Codex                       | $1.07   | $8.50   | $0.107           | -                  |
 | GPT 5 Nano                        | $0.05   | $0.40   | $0.005           | -                  |

+**GPT 5.6 Sol:** Los precios mostrados incluyen un 50 % de descuento hasta el 18 de septiembre de 2026.
+
 **DeepSeek V4 Flash / Pro:** Las horas Peak son 01:00-04:00 y 06:00-10:00 UTC; todas las demás horas son Off-Peak. [Más información](https://api-docs.deepseek.com/quick_start/pricing/).

 Puede que notes [modelos de bajo costo](/docs/config/#models), como Haiku, Nano o Flash, en tu historial de uso. OpenCode usa estos modelos para generar títulos de sesiones.
@@ -68,6 +68,7 @@ La liste actuelle des modèles comprend :
 - **Qwen3.6 Plus**
 - **DeepSeek V4 Pro**
 - **DeepSeek V4 Flash**
+- **DeepSeek V4 Flash Vision Exp**
 - **Hy3**
 - **Ox Alpha Free** (pour une durée limitée)

@@ -87,29 +88,30 @@ Les limites sont définies en valeur monétaire (dollars). Cela signifie que vot

 Le tableau ci-dessous fournit une estimation du nombre de requêtes basée sur des modèles d'utilisation typiques de Go :

-| Model                      | requêtes par 5 heures | requêtes par semaine | requêtes par mois |
-| -------------------------- | --------------------- | -------------------- | ----------------- |
-| Grok 4.5                   | 120                   | 300                  | 600               |
-| GPT 5.6 Luna               | 2,050                 | 5,100                | 10,250            |
-| GLM-5.3                    | 220                   | 540                  | 1,080             |
-| GLM-5.2                    | 880                   | 2,150                | 4,300             |
-| GLM-5.1                    | 880                   | 2,150                | 4,300             |
-| Kimi K3                    | 110                   | 250                  | 490               |
-| Kimi K2.7 Code             | 1,350                 | 3,380                | 6,750             |
-| Kimi K2.6                  | 1,150                 | 2,880                | 5,750             |
-| MiMo-V2.5                  | 30,100                | 75,200               | 150,400           |
-| MiMo-V2.5-Pro              | 3,250                 | 8,150                | 16,300            |
-| MiniMax M3                 | 3,200                 | 8,000                | 16,000            |
-| MiniMax M2.7               | 3,400                 | 8,500                | 17,000            |
-| Muse Spark 1.2 Contributor | 45,300                | 113,300              | 226,600           |
-| Qwen3.8 Max                | 160                   | 400                  | 810               |
-| Qwen3.7 Max                | 340                   | 840                  | 1,690             |
-| Qwen3.7 Plus               | 4,300                 | 10,800               | 21,600            |
-| Qwen3.6 Plus               | 3,300                 | 8,200                | 16,300            |
-| DeepSeek V4 Pro            | 1,050                 | 2,600                | 5,200             |
-| DeepSeek V4 Flash          | 7,600                 | 18,900               | 37,800            |
-| Hy3                        | 4,300                 | 10,750               | 21,500            |
-| Ox Alpha Free              | -                     | -                    | -                 |
+| Model                        | requêtes par 5 heures | requêtes par semaine | requêtes par mois |
+| ---------------------------- | --------------------- | -------------------- | ----------------- |
+| Grok 4.5                     | 120                   | 300                  | 600               |
+| GPT 5.6 Luna                 | 2,050                 | 5,100                | 10,250            |
+| GLM-5.3                      | 220                   | 540                  | 1,080             |
+| GLM-5.2                      | 880                   | 2,150                | 4,300             |
+| GLM-5.1                      | 880                   | 2,150                | 4,300             |
+| Kimi K3                      | 110                   | 250                  | 490               |
+| Kimi K2.7 Code               | 1,350                 | 3,380                | 6,750             |
+| Kimi K2.6                    | 1,150                 | 2,880                | 5,750             |
+| MiMo-V2.5                    | 30,100                | 75,200               | 150,400           |
+| MiMo-V2.5-Pro                | 3,250                 | 8,150                | 16,300            |
+| MiniMax M3                   | 3,200                 | 8,000                | 16,000            |
+| MiniMax M2.7                 | 3,400                 | 8,500                | 17,000            |
+| Muse Spark 1.2 Contributor   | 45,300                | 113,300              | 226,600           |
+| Qwen3.8 Max                  | 160                   | 400                  | 810               |
+| Qwen3.7 Max                  | 340                   | 840                  | 1,690             |
+| Qwen3.7 Plus                 | 4,300                 | 10,800               | 21,600            |
+| Qwen3.6 Plus                 | 3,300                 | 8,200                | 16,300            |
+| DeepSeek V4 Pro              | 1,050                 | 2,600                | 5,200             |
+| DeepSeek V4 Flash            | 7,600                 | 18,900               | 37,800            |
+| DeepSeek V4 Flash Vision Exp | 3,800                 | 9,450                | 18,900            |
+| Hy3                          | 4,300                 | 10,750               | 21,500            |
+| Ox Alpha Free                | -                     | -                    | -                 |

 Les estimations sont basées sur les schémas de requêtes observés :

@@ -120,6 +122,7 @@ Les estimations sont basées sur les schémas de requêtes observés :
 - Kimi K2.7/K2.6 — 870 tokens en entrée, 55,000 en cache, 200 tokens en sortie par requête
 - DeepSeek V4 Pro — 750 tokens en entrée, 82,000 en cache, 290 tokens en sortie par requête
 - DeepSeek V4 Flash — 410 tokens en entrée, 71,300 en cache, 310 tokens en sortie par requête
+- DeepSeek V4 Flash Vision Exp — 410 tokens en entrée, 71,300 en cache, 310 tokens en sortie par requête
 - MiniMax M3 — 510 tokens en entrée, 56,000 en cache, 190 tokens en sortie par requête
 - MiniMax M2.7 — 300 tokens en entrée, 55,000 en cache, 125 tokens en sortie par requête
 - Muse Spark 1.2 Contributor — 620 tokens en entrée, 71,400 en cache, 300 tokens en sortie par requête
@@ -133,37 +136,41 @@ Les estimations sont basées sur les schémas de requêtes observés :

 Les estimations sont également basées sur les prix suivants par 1M tokens et sur l'utilisation mensuelle incluse avec chaque modèle :

-| Modèle                       | Input  | Output | Cached Read | Cached Write | Utilisation |
-| ---------------------------- | ------ | ------ | ----------- | ------------ | ----------- |
-| Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15         |
-| GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15         |
-| GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15         |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15         |
-| GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60         |
-| GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60         |
-| Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15         |
-| Kimi K2.7 Code               | $0.95  | $4.00  | $0.19       | -            | $60         |
-| Kimi K2.6                    | $0.95  | $4.00  | $0.16       | -            | $60         |
-| MiMo V2.5                    | $0.14  | $0.28  | $0.0028     | -            | $60         |
-| MiMo V2.5 Pro                | $0.435 | $0.87  | $0.003625   | -            | $15         |
-| MiniMax M3                   | $0.30  | $1.20  | $0.06       | -            | $60         |
-| MiniMax M2.7                 | $0.30  | $1.20  | $0.06       | $0.375       | $60         |
-| MiniMax M2.5                 | $0.30  | $1.20  | $0.06       | $0.375       | $60         |
-| Muse Spark 1.2 Contributor   | $0.10  | $0.20  | $0.002      | -            | $60         |
-| Qwen3.8 Max                  | $2.00  | $6.00  | $0.25       | $2.50        | $15         |
-| Qwen3.7 Max                  | $2.50  | $7.50  | $0.50       | $3.125       | $60         |
-| Qwen3.7 Plus (≤ 256K tokens) | $0.40  | $1.60  | $0.04       | $0.50        | $60         |
-| Qwen3.7 Plus (> 256K tokens) | $1.20  | $4.80  | $0.12       | $1.50        | $60         |
-| Qwen3.6 Plus (≤ 256K tokens) | $0.50  | $3.00  | $0.05       | $0.625       | $60         |
-| Qwen3.6 Plus (> 256K tokens) | $2.00  | $6.00  | $0.20       | $2.50        | $60         |
-| DeepSeek V4 Pro (Off-Peak)   | $0.66  | $1.98  | $0.022      | -            | $15         |
-| DeepSeek V4 Pro (Peak)       | $1.32  | $3.96  | $0.044      | -            | $15         |
-| DeepSeek V4 Flash (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $30         |
-| DeepSeek V4 Flash (Peak)     | $0.44  | $1.32  | $0.014      | -            | $30         |
-| Hy3                          | $0.14  | $0.58  | $0.035      | -            | $60         |
-| Ox Alpha Free                | -      | -      | -           | -            | -           |
+| Modèle                                  | Input  | Output | Cached Read | Cached Write | Utilisation |
+| --------------------------------------- | ------ | ------ | ----------- | ------------ | ----------- |
+| Grok 4.5                                | $2.00  | $6.00  | $0.30       | -            | $15         |
+| GPT 5.6 Luna (≤ 272K tokens)            | $0.20  | $1.20  | $0.02       | $0.25        | $15         |
+| GPT 5.6 Luna (> 272K tokens)            | $0.40  | $1.80  | $0.04       | $0.50        | $15         |
+| GLM-5.3                                 | $1.40  | $4.40  | $0.26       | -            | $15         |
+| GLM-5.2                                 | $1.40  | $4.40  | $0.26       | -            | $60         |
+| GLM-5.1                                 | $1.40  | $4.40  | $0.26       | -            | $60         |
+| Kimi K3                                 | $3.00  | $15.00 | $0.30       | -            | $15         |
+| Kimi K2.7 Code                          | $0.95  | $4.00  | $0.19       | -            | $60         |
+| Kimi K2.6                               | $0.95  | $4.00  | $0.16       | -            | $60         |
+| MiMo V2.5                               | $0.14  | $0.28  | $0.0028     | -            | $60         |
+| MiMo V2.5 Pro                           | $0.435 | $0.87  | $0.003625   | -            | $15         |
+| MiniMax M3                              | $0.30  | $1.20  | $0.06       | -            | $60         |
+| MiniMax M2.7                            | $0.30  | $1.20  | $0.06       | $0.375       | $60         |
+| MiniMax M2.5                            | $0.30  | $1.20  | $0.06       | $0.375       | $60         |
+| Muse Spark 1.2 Contributor              | $0.10  | $0.20  | $0.002      | -            | $60         |
+| Qwen3.8 Max                             | $2.00  | $6.00  | $0.25       | $2.50        | $15         |
+| Qwen3.7 Max                             | $2.50  | $7.50  | $0.50       | $3.125       | $60         |
+| Qwen3.7 Plus (≤ 256K tokens)            | $0.40  | $1.60  | $0.04       | $0.50        | $60         |
+| Qwen3.7 Plus (> 256K tokens)            | $1.20  | $4.80  | $0.12       | $1.50        | $60         |
+| Qwen3.6 Plus (≤ 256K tokens)            | $0.50  | $3.00  | $0.05       | $0.625       | $60         |
+| Qwen3.6 Plus (> 256K tokens)            | $2.00  | $6.00  | $0.20       | $2.50        | $60         |
+| DeepSeek V4 Pro (Off-Peak)              | $0.66  | $1.98  | $0.022      | -            | $15         |
+| DeepSeek V4 Pro (Peak)                  | $1.32  | $3.96  | $0.044      | -            | $15         |
+| DeepSeek V4 Flash (Off-Peak)            | $0.22  | $0.66  | $0.007      | -            | $30         |
+| DeepSeek V4 Flash (Peak)                | $0.44  | $1.32  | $0.014      | -            | $30         |
+| DeepSeek V4 Flash Vision Exp (Off-Peak) | $0.22  | $0.66  | $0.007      | -            | $15         |
+| DeepSeek V4 Flash Vision Exp (Peak)     | $0.44  | $1.32  | $0.014      | -            | $15         |
+| Hy3                                     | $0.14  | $0.58  | $0.035      | -            | $60         |
+| Ox Alpha Free                           | -      | -      | -           | -            | -           |

-**DeepSeek V4 Flash / Pro:** Les heures Peak sont 01:00-04:00 et 06:00-10:00 UTC ; toutes les autres heures sont Off-Peak. [En savoir plus](https://api-docs.deepseek.com/quick_start/pricing/).
+**DeepSeek V4 Flash / V4 Flash Vision Exp / V4 Pro:** Les heures Peak sont 01:00-04:00 et 06:00-10:00 UTC ; toutes les autres heures sont Off-Peak. [En savoir plus](https://api-docs.deepseek.com/quick_start/pricing/).
+
+**DeepSeek V4 Flash Vision Exp:** Les images sont converties en tokens selon leurs dimensions et facturées comme tokens d’entrée avec les tokens de texte. [En savoir plus](https://api-docs.deepseek.com/quick_start/pricing/).

 **Ox Alpha Free:** Gratuit pour une durée limitée.

@@ -199,30 +206,31 @@ Pour ces modèles, vous obtenez tout de même un peu plus que si vous payiez dir

 Vous pouvez également accéder aux modèles Go via les points de terminaison d'API suivants.

-| Modèle                     | ID de modèle               | Point de terminaison                             | Package AI SDK              |
-| -------------------------- | -------------------------- | ------------------------------------------------ | --------------------------- |
-| Grok 4.5                   | grok-4.5                   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GPT 5.6 Luna               | gpt-5.6-luna               | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3                    | glm-5.3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.2                    | glm-5.2                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| GLM-5.1                    | glm-5.1                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K3                    | kimi-k3                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code             | kimi-k2.7-code             | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Kimi K2.6                  | kimi-k2.6                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Pro            | deepseek-v4-pro            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash          | deepseek-v4-flash          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5                  | mimo-v2.5                  | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro              | mimo-v2.5-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiniMax M3                 | minimax-m3                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.7               | minimax-m2.7               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| MiniMax M2.5               | minimax-m2.5               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Muse Spark 1.2 Contributor | muse-spark-1.2-contributor | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| Qwen3.8 Max                | qwen3.8-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Max                | qwen3.7-max                | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus               | qwen3.7-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus               | qwen3.6-plus               | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
-| Hy3                        | hy3                        | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| Ox Alpha Free              | x-preview-f-free           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Modèle                       | ID de modèle                 | Point de terminaison                             | Package AI SDK              |
+| ---------------------------- | ---------------------------- | ------------------------------------------------ | --------------------------- |
+| Grok 4.5                     | grok-4.5                     | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                 | gpt-5.6-luna                 | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| GLM-5.3                      | glm-5.3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.2                      | glm-5.2                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| GLM-5.1                      | glm-5.1                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K3                      | kimi-k3                      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code               | kimi-k2.7-code               | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                    | kimi-k2.6                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Pro              | deepseek-v4-pro              | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash            | deepseek-v4-flash            | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Vision Exp | deepseek-v4-flash-vision-exp | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5                    | mimo-v2.5                    | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro                | mimo-v2.5-pro                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiniMax M3                   | minimax-m3                   | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.7                 | minimax-m2.7                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| MiniMax M2.5                 | minimax-m2.5                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Muse Spark 1.2 Contributor   | muse-spark-1.2-contributor   | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
+| Qwen3.8 Max                  | qwen3.8-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Max                  | qwen3.7-max                  | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                 | qwen3.7-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                 | qwen3.6-plus                 | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
+| Hy3                          | hy3                          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| Ox Alpha Free                | ox-alpha-free                | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |

 L'[ID de modèle](/docs/config/#models) dans votre configuration OpenCode utilise le format `opencode-go/<model-id>`. Par exemple, pour Kimi K3, vous utiliseriez `opencode-go/kimi-k3` dans votre configuration.

@@ -240,29 +248,30 @@ https://opencode.ai/zen/go/v1/models

 ## Confidentialité

-| Modèle                     | Entraînement des modèles | Conservation des données |
-| -------------------------- | ------------------------ | ------------------------ |
-| Grok 4.5                   | Non utilisé              | 30 jours                 |
-| GPT 5.6 Luna               | Non utilisé              | 30 jours                 |
-| GLM-5.3                    | Non utilisé              | 0 jour                   |
-| GLM-5.2                    | Non utilisé              | 0 jour                   |
-| GLM-5.1                    | Non utilisé              | 0 jour                   |
-| Kimi K3                    | Non utilisé              | 0 jour                   |
-| Kimi K2.7 Code             | Non utilisé              | 0 jour                   |
-| Kimi K2.6                  | Non utilisé              | 0 jour                   |
-| MiMo-V2.5-Pro              | Non utilisé              | 0 jour                   |
-| MiMo-V2.5                  | Non utilisé              | 0 jour                   |
-| Qwen3.8 Max                | Non utilisé              | 0 jour                   |
-| Qwen3.7 Max                | Non utilisé              | 0 jour                   |
-| Qwen3.7 Plus               | Non utilisé              | 0 jour                   |
-| Qwen3.6 Plus               | Non utilisé              | 0 jour                   |
-| MiniMax M3                 | Non utilisé              | 0 jour                   |
-| MiniMax M2.7               | Non utilisé              | 0 jour                   |
-| Muse Spark 1.2 Contributor | Oui                      | Pas de ZDR               |
-| DeepSeek V4 Pro            | Non utilisé              | 0 jour                   |
-| DeepSeek V4 Flash          | Non utilisé              | 0 jour                   |
-| Hy3                        | Non utilisé              | 0 jour                   |
-| Ox Alpha Free              | Non utilisé              | 0 jour                   |
+| Modèle                       | Entraînement des modèles | Conservation des données |
+| ---------------------------- | ------------------------ | ------------------------ |
+| Grok 4.5                     | Non utilisé              | 30 jours                 |
+| GPT 5.6 Luna                 | Non utilisé              | 30 jours                 |
+| GLM-5.3                      | Non utilisé              | 0 jour                   |
+| GLM-5.2                      | Non utilisé              | 0 jour                   |
+| GLM-5.1                      | Non utilisé              | 0 jour                   |
+| Kimi K3                      | Non utilisé              | 0 jour                   |
+| Kimi K2.7 Code               | Non utilisé              | 0 jour                   |
+| Kimi K2.6                    | Non utilisé              | 0 jour                   |
+| MiMo-V2.5-Pro                | Non utilisé              | 0 jour                   |
+| MiMo-V2.5                    | Non utilisé              | 0 jour                   |
+| Qwen3.8 Max                  | Non utilisé              | 0 jour                   |
+| Qwen3.7 Max                  | Non utilisé              | 0 jour                   |
+| Qwen3.7 Plus                 | Non utilisé              | 0 jour                   |
+| Qwen3.6 Plus                 | Non utilisé              | 0 jour                   |
+| MiniMax M3                   | Non utilisé              | 0 jour                   |
+| MiniMax M2.7                 | Non utilisé              | 0 jour                   |
+| Muse Spark 1.2 Contributor   | Oui                      | Pas de ZDR               |
+| DeepSeek V4 Pro              | Non utilisé              | 0 jour                   |
+| DeepSeek V4 Flash            | Non utilisé              | 0 jour                   |
+| DeepSeek V4 Flash Vision Exp | Non utilisé              | 0 jour                   |
+| Hy3                          | Non utilisé              | 0 jour                   |
+| Ox Alpha Free                | Non utilisé              | 0 jour                   |

 - **Grok 4.5:** Le ZDR désactive d’importantes fonctionnalités API qui dépendent des données stockées, notamment Responses API avec état, Files and Collections et Batch API. [En savoir plus](https://docs.x.ai/developers/faq/security#what-is-zero-data-retention-zdr).
 - **GPT 5.6 Luna:** Des journaux de surveillance des abus sont générés pour toute utilisation des fonctionnalités API et conservés pendant un maximum de 30 jours. [En savoir plus](https://developers.openai.com/api/docs/guides/your-data#data-retention-controls-for-abuse-monitoring).
@@ -184,8 +184,8 @@ Nous prenons en charge un modèle de paiement à l'utilisation. Vous trouverez c
 | Grok 4.5 (> 200K tokens)          | $4.00  | $12.00  | $0.60       | -            |
 | Grok Build 0.1                    | $1.00  | $2.00   | $0.20       | -            |
 | Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
-| GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
-| GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
+| GPT 5.6 Sol (≤ 272K tokens)       | $2.00  | $10.00  | $0.20       | $2.50        |
+| GPT 5.6 Sol (> 272K tokens)       | $4.00  | $15.00  | $0.40       | $5.00        |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
 | GPT 5.6 Terra (> 272K tokens)     | $4.00  | $18.00  | $0.40       | $5.00        |
 | GPT 5.6 Luna (≤ 272K tokens)      | $0.20  | $1.20   | $0.02       | $0.25        |
@@ -210,6 +210,8 @@ Nous prenons en charge un modèle de paiement à l'utilisation. Vous trouverez c
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |

+**GPT 5.6 Sol:** Les prix affichés incluent une réduction de 50 % jusqu’au 18 septembre 2026.
+
 **DeepSeek V4 Flash / Pro:** Les heures Peak sont 01:00-04:00 et 06:00-10:00 UTC ; toutes les autres heures sont Off-Peak. [En savoir plus](https://api-docs.deepseek.com/quick_start/pricing/).

 Vous remarquerez peut-être des [modèles à faible coût](/docs/config/#models), tels que Haiku, Nano ou Flash, dans votre historique d'utilisation. OpenCode utilise ces modèles pour générer les titres des sessions.
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
opencode-agent[bot]	03bba464d4	chore: generate	2026-08-23 17:15:36 +00:00
Dax	63a883a4f7	refactor(console): stream zen bodies without modifiers (#44472 )	2026-08-23 13:14:18 -04:00
Dax	b3bad6b581	fix(console): revert scanned zen request streaming (#44470 )	2026-08-23 13:03:20 -04:00
opencode-agent[bot]	ca10088bdf	chore: generate	2026-08-23 16:51:36 +00:00
Dax	dd3f915956	fix(console): release streamed zen prefixes (#44465 )	2026-08-23 12:50:18 -04:00
opencode-agent[bot]	c11c41bd86	chore: generate	2026-08-23 16:48:45 +00:00
Dax	fa117558ee	fix(console): scan zen model before streaming (#44463 )	2026-08-23 12:47:29 -04:00
Frank	17b47301ba	feat(console): add unblock workspace action	2026-08-23 12:46:02 -04:00
Dax	3d31e4bcec	fix(console): revert streamed zen request bodies (#44444 )	2026-08-23 11:54:07 -04:00
opencode-agent[bot]	bb72277407	chore: generate	2026-08-23 15:19:54 +00:00
Dax	32c3637da0	fix(console): stream zen request bodies (#44429 )	2026-08-23 11:18:38 -04:00
Dax Raad	9d466cd849	fix(stats): normalize model casing	2026-08-23 10:33:11 -04:00
opencode-agent[bot]	e3bd6e0947	chore: generate	2026-08-23 14:00:45 +00:00
Dax	dc13c6bb3d	fix(console): reduce zen request memory (#44403 )	2026-08-23 09:59:29 -04:00
Brendan Allan	3a31c4ea80	fix(app): keep model provider headers visible (#44115 )	2026-08-22 20:50:02 +08:00
Kit Langton	e00890c672	fix: resolve console device URLs (#44029 )	2026-08-22 00:25:56 +00:00
joelstucki-taulia	3a4c253969	fix(provider): guard textVerbosity injection for @ai-sdk/openai-compatible providers (#43915 ) Co-authored-by: Joel Stucki <joel.stucki@example.com>	2026-08-21 22:19:33 +00:00
Adam	34a83b27c3	fix(stats): improve chart tooltip truncation	2026-08-21 17:13:13 -05:00
opencode-agent[bot]	ff3ef6e3e6	chore: update nix node_modules hashes	2026-08-21 21:54:44 +00:00
opencode-agent[bot]	e2ec62d073	fix: bump Amazon Bedrock provider (#43909 ) Co-authored-by: neriousy <34747899+neriousy@users.noreply.github.com>	2026-08-21 23:39:17 +02:00
opencode-agent[bot]	bcf1103a8c	docs(zen): update GPT 5.6 Sol pricing (#43986 ) Co-authored-by: Slickstef11 <98915060+Slickstef11@users.noreply.github.com>	2026-08-21 17:00:03 -04:00
Dax Raad	9cb6fb6529	fix(console): cache desktop downloads	2026-08-21 16:58:52 -04:00
opencode-agent[bot]	1b937c860b	test(opencode): align unknown finish coverage (#43895 ) Co-authored-by: thdxr <826656+thdxr@users.noreply.github.com>	2026-08-21 10:56:50 -04:00
opencode	ad0bb6d9a3	sync release versions for v1.18.21	2026-08-21 14:51:08 +00:00
opencode-agent[bot]	487d584402	docs(zen): reflect GPT 5.6 Sol discount (#43874 )	2026-08-21 22:49:25 +08:00
opencode-agent[bot]	57fa34f235	fix(opencode): continue unknown finish responses (#43892 ) Co-authored-by: rekram1-node <rekram1-node@users.noreply.github.com>	2026-08-21 10:30:15 -04:00
Adam	0af9dd61d2	fix(stats): merge renamed model data (#43883 )	2026-08-21 08:38:31 -05:00
opencode-agent[bot]	8ecd4c21bf	chore: generate	2026-08-21 13:24:26 +00:00
opencode-agent[bot]	361a71ffad	fix(opencode): route Vertex multi-regions through REP (#42648 ) Co-authored-by: Aiden Cline <rekram1-node@users.noreply.github.com> Co-authored-by: Filip <34747899+neriousy@users.noreply.github.com>	2026-08-21 15:23:05 +02:00
opencode-agent[bot]	fa8f17014c	chore: generate	2026-08-21 12:58:43 +00:00
Jack	813e6f3cec	docs(go): add DeepSeek vision model (#43866 )	2026-08-21 20:57:25 +08:00
Jack	ba72a6ff2b	fix(go): align Ox Alpha model ID (#43837 )	2026-08-21 17:03:23 +08:00
Brendan Allan	b6a1f95858	fix(app): preserve file search results while loading (#43836 )	2026-08-21 08:52:45 +00:00
Nathan Thomassin	f357e70779	fix(app): register archive session command in both layouts (#41741 )	2026-08-21 08:28:40 +00:00
opencode	a9cac91d60	sync release versions for v1.18.20	2026-08-21 08:09:28 +00:00
Aiden Cline	35fe5b7212	fix(opencode): surface subagent tool errors (#43821 )	2026-08-21 02:44:44 -05:00
opencode-agent[bot]	62cb3f77bd	test(opencode): remove flaky subagent test (#43819 ) Co-authored-by: Aiden <rekram1-node@users.noreply.github.com>	2026-08-21 02:35:25 -05:00
opencode-agent[bot]	e0b9e68a68	fix(opencode): retry raw network finish errors (#43813 ) Co-authored-by: Aiden <rekram1-node@users.noreply.github.com>	2026-08-21 02:17:35 -05:00
opencode-agent[bot]	1e4c153aed	chore: generate	2026-08-21 07:07:33 +00:00
Jack	c9dc1a1b3d	feat(go): add Ox Alpha to usage graph (#43809 )	2026-08-21 15:06:23 +08:00