feat(opencode): cache-friendly compaction via primary loop request path

Default compaction no longer uses a dedicated compaction agent. It inherits the session's active agent and delegates request assembly to the primary loop's shared envelope (same system prompt, instructions, skills, MCP context, and tool definitions), replaying the typed message head and appending the summary instruction as the final user message. This keeps the compaction request prefix-identical to the preceding turn so provider prompt/KV caches are reused instead of fully re-prefilled. An explicitly configured agent.compaction preserves the legacy behavior exactly: dedicated hidden agent, optional model override, and the serialized-transcript summary request.
docs(go): use responses API for Grok 4.5 (#42373 )
2026-08-14 07:19:49 -04:00 · 2026-08-13 23:10:07 -05:00 · 2026-08-14 01:28:55 +08:00 · 2026-08-13 13:27:41 -04:00 · 2026-08-13 16:51:43 +00:00 · 2026-08-13 16:30:12 +00:00
136 changed files with 3152 additions and 1767 deletions
@@ -10,6 +10,7 @@ kitlangton
 kommander
 ludvigrask
 MrMushrooooom
+neriousy
 nexxeln
 R44VC0RP
 rekram1-node
@@ -35,6 +35,7 @@ jobs:

      - run: bun sst deploy --stage=${{ github.ref_name }}
        env:
+          GITHUB_TOKEN: ${{ github.token }}
          CLOUDFLARE_API_TOKEN: ${{ secrets.CLOUDFLARE_API_TOKEN }}
          PLANETSCALE_SERVICE_TOKEN_NAME: ${{ secrets.PLANETSCALE_SERVICE_TOKEN_NAME }}
          PLANETSCALE_SERVICE_TOKEN: ${{ secrets.PLANETSCALE_SERVICE_TOKEN }}
@@ -4,7 +4,7 @@ import { tool } from "@opencode-ai/plugin"
 const TEAM = {
  tui: ["kommander", "simonklee"],
  desktop_web: ["Hona", "Brendonovich"],
-  core: ["jlongster", "rekram1-node", "nexxeln", "kitlangton"],
+  core: ["jlongster", "rekram1-node", "neriousy", "nexxeln", "kitlangton"],
  inference: ["fwang", "MrMushrooooom", "starptech"],
  windows: ["Hona"],
 } as const
@@ -29,7 +29,7 @@
    },
    "packages/app": {
      "name": "@opencode-ai/app",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@corvu/drawer": "catalog:",
        "@dnd-kit/abstract": "0.5.0",
@@ -96,7 +96,7 @@
    },
    "packages/cli": {
      "name": "@opencode-ai/cli",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "bin": {
        "lildax": "./bin/lildax.cjs",
      },
@@ -144,7 +144,7 @@
    },
    "packages/codemode": {
      "name": "@opencode-ai/codemode",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "acorn": "8.15.0",
        "effect": "catalog:",
@@ -158,7 +158,7 @@
    },
    "packages/console/app": {
      "name": "@opencode-ai/console-app",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@cloudflare/vite-plugin": "1.15.2",
        "@opencode-ai/console-core": "workspace:*",
@@ -287,7 +287,7 @@
    },
    "packages/core": {
      "name": "@opencode-ai/core",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "bin": {
        "opencode": "./bin/opencode",
      },
@@ -381,7 +381,7 @@
    },
    "packages/desktop": {
      "name": "@opencode-ai/desktop",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "drizzle-orm": "catalog:",
        "effect": "catalog:",
@@ -449,7 +449,7 @@
    },
    "packages/effect-sqlite-node": {
      "name": "@opencode-ai/effect-sqlite-node",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "effect": "catalog:",
      },
@@ -461,7 +461,7 @@
    },
    "packages/enterprise": {
      "name": "@opencode-ai/enterprise",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@hono/standard-validator": "catalog:",
        "@opencode-ai/core": "workspace:*",
@@ -493,7 +493,7 @@
    },
    "packages/function": {
      "name": "@opencode-ai/function",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@octokit/auth-app": "8.0.1",
        "@octokit/rest": "catalog:",
@@ -509,7 +509,7 @@
    },
    "packages/http-recorder": {
      "name": "@opencode-ai/http-recorder",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@opencode-ai/schema": "workspace:*",
        "@smithy/eventstream-codec": "4.2.14",
@@ -559,7 +559,7 @@
    },
    "packages/opencode": {
      "name": "opencode",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "bin": {
        "opencode": "./bin/opencode",
      },
@@ -690,7 +690,7 @@
    },
    "packages/plugin": {
      "name": "@opencode-ai/plugin",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@ai-sdk/provider": "3.0.8",
        "@opencode-ai/sdk": "workspace:*",
@@ -766,7 +766,7 @@
    },
    "packages/sdk/js": {
      "name": "@opencode-ai/sdk",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "cross-spawn": "catalog:",
      },
@@ -781,7 +781,7 @@
    },
    "packages/server": {
      "name": "@opencode-ai/server",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@opencode-ai/core": "workspace:*",
        "@opencode-ai/protocol": "workspace:*",
@@ -796,7 +796,7 @@
    },
    "packages/session-ui": {
      "name": "@opencode-ai/session-ui",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@opencode-ai/sdk": "workspace:*",
        "@slack/bolt": "^3.17.1",
@@ -849,7 +849,7 @@
    },
    "packages/stats/app": {
      "name": "@opencode-ai/stats-app",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@ibm/plex": "6.4.1",
        "@kobalte/core": "catalog:",
@@ -883,7 +883,7 @@
    },
    "packages/stats/core": {
      "name": "@opencode-ai/stats-core",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "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": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@aws-sdk/client-firehose": "3.933.0",
        "@effect/platform-node": "catalog:",
@@ -944,7 +944,7 @@
    },
    "packages/tui": {
      "name": "@opencode-ai/tui",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@opencode-ai/core": "workspace:*",
        "@opencode-ai/plugin": "workspace:*",
@@ -971,7 +971,7 @@
    },
    "packages/ui": {
      "name": "@opencode-ai/ui",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@kobalte/core": "catalog:",
        "@pierre/diffs": "catalog:",
@@ -1022,7 +1022,7 @@
    },
    "packages/web": {
      "name": "@opencode-ai/web",
-      "version": "1.18.14",
+      "version": "1.18.18",
      "dependencies": {
        "@astrojs/cloudflare": "12.6.3",
        "@astrojs/markdown-remark": "6.3.1",
@@ -1075,6 +1075,7 @@
    "@silvia-odwyer/photon-node@0.3.4": "patches/@silvia-odwyer%2Fphoton-node@0.3.4.patch",
    "@npmcli/agent@4.0.2": "patches/@npmcli%2Fagent@4.0.2.patch",
    "solid-js@1.9.10": "patches/solid-js@1.9.10.patch",
+    "@ai-sdk/groq@3.0.31": "patches/@ai-sdk%2Fgroq@3.0.31.patch",
    "@ai-sdk/google@3.0.73": "patches/@ai-sdk%2Fgoogle@3.0.73.patch",
    "pacote@21.5.0": "patches/pacote@21.5.0.patch",
    "@dnd-kit/dom@0.5.0": "patches/@dnd-kit%2Fdom@0.5.0.patch",
@@ -181,6 +181,16 @@ const statsSyncConfig = new sst.Linkable("StatsSyncConfig", {
  },
 })

+const r2SqlAuthToken = new sst.Secret("R2SqlAuthToken")
+const r2Sql = new sst.Linkable("R2Sql", {
+  properties: {
+    accountId: "15d29c8639fd3733b1b5486a2acfd968",
+    bucket: `platform-${$app.stage}-lake`,
+    namespace: "inference",
+    table: "generation",
+  },
+})
+
 export const statSync = new sst.aws.Service("StatsSyncService", {
  cluster: lakeCluster,
  architecture: "arm64",
@@ -193,7 +203,9 @@ export const statSync = new sst.aws.Service("StatsSyncService", {
    dockerfile: "packages/stats/server/Dockerfile",
  },
  command: ["bun", "src/stat-sync.ts"],
-  link: [database, inferenceEvent, statsSyncConfig],
+  // Keep the legacy Athena link and IAM permissions during the first R2-backed
+  // release so reverting the application code remains a one-deploy rollback.
+  link: [database, inferenceEvent, r2Sql, r2SqlAuthToken, statsSyncConfig],
  permissions: lakeQueryPermissions,
  scaling: {
    min: 1,
@@ -1,8 +1,8 @@
 {
  "nodeModules": {
-    "x86_64-linux": "sha256-uduwrM143NDSc+tXsi4lVVfoMll2a3BDHRUjuO7GB68=",
-    "aarch64-linux": "sha256-6DUda78XdXY6DP86lIUkweSjys3iG4Y4mo1PiaNuXbg=",
-    "aarch64-darwin": "sha256-AkJwfLULLZVwwz+XU1QcFUZoIS7oVPCn+n/MXEaxrqE=",
-    "x86_64-darwin": "sha256-hAxKGdiITTxQ2uujQt6prNjo3NxGAMMeo+9HlMWK6GU="
+    "x86_64-linux": "sha256-TNwKfqxD83UpZuCKN8FdEWN+CcQUP9CkCQSLGNqR/sA=",
+    "aarch64-linux": "sha256-qzvOJZzmq2QhlauElw8GwgQnCPHdhexI52L0md5zrxQ=",
+    "aarch64-darwin": "sha256-ZzoyLayOFfcYUAg35ZbZ2WapxDdd9IUWqy2xkxZH4QM=",
+    "x86_64-darwin": "sha256-maP/qLeaC3q8VcmNIPyIKlnplxFXJ7ULho3v21/16Mw="
  }
 }
@@ -159,6 +159,7 @@
    "@modelcontextprotocol/sdk@1.29.0": "patches/@modelcontextprotocol%2Fsdk@1.29.0.patch",
    "effect@4.0.0-beta.83": "patches/effect@4.0.0-beta.83.patch",
    "@tanstack/virtual-core@3.17.3": "patches/@tanstack%2Fvirtual-core@3.17.3.patch",
-    "@ai-sdk/openai-compatible@2.0.41": "patches/@ai-sdk%2Fopenai-compatible@2.0.41.patch"
+    "@ai-sdk/openai-compatible@2.0.41": "patches/@ai-sdk%2Fopenai-compatible@2.0.41.patch",
+    "@ai-sdk/groq@3.0.31": "patches/@ai-sdk%2Fgroq@3.0.31.patch"
  }
 }
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/app",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "description": "",
  "type": "module",
  "exports": {
@@ -139,6 +139,7 @@ test("resolves directory autocomplete from the current browser root", async () =
          directories.push(input.location?.directory ?? "")
          return Promise.resolve({ data: [] })
        },
+        list: () => Promise.resolve({ data: [] }),
      },
    },
  } as unknown as Parameters<typeof createDirectorySearch>[0]["sdk"]
@@ -152,6 +153,67 @@ test("resolves directory autocomplete from the current browser root", async () =
  expect(directories).toEqual(["/repo", "/repo/src"])
 })

+test("keeps indexed directory results for servers that support empty search", async () => {
+  const sdk = {
+    api: {
+      file: {
+        find: () => Promise.resolve({ data: [{ path: "projects/", type: "directory" }] }),
+        list: () => Promise.reject(new Error("listing should not run when search returns results")),
+      },
+    },
+  } as unknown as Parameters<typeof createDirectorySearch>[0]["sdk"]
+  const search = createDirectorySearch({ sdk, home: () => "/home/luke", base: () => "/home/luke" })
+
+  expect(await search("")).toEqual(["/home/luke/projects"])
+})
+
+test("lists the default directory when empty search is unsupported", async () => {
+  const calls: string[] = []
+  const directories = Array.from({ length: 60 }, (_, index) => ({
+    path: `project-${index}/`,
+    type: "directory" as const,
+  }))
+  const sdk = {
+    api: {
+      file: {
+        find: () => Promise.resolve({ data: [] }),
+        list: (input: { location?: { directory?: string } }) => {
+          calls.push(input.location?.directory ?? "")
+          return Promise.resolve({
+            data: [...directories, { path: "README.md", type: "file" }],
+          })
+        },
+      },
+    },
+  } as unknown as Parameters<typeof createDirectorySearch>[0]["sdk"]
+  const search = createDirectorySearch({ sdk, home: () => "/home/luke", base: () => "/home/luke" })
+
+  const results = await search("")
+  expect(results).toHaveLength(60)
+  expect(results.at(-1)).toBe("/home/luke/project-59")
+  expect(calls).toEqual(["/home/luke"])
+})
+
+test("matches the default directory listing when typed search is unsupported", async () => {
+  const sdk = {
+    api: {
+      file: {
+        find: () => Promise.resolve({ data: [] }),
+        list: () =>
+          Promise.resolve({
+            data: [
+              { path: "Documents/", type: "directory" },
+              { path: "Downloads/", type: "directory" },
+            ],
+          }),
+      },
+    },
+  } as unknown as Parameters<typeof createDirectorySearch>[0]["sdk"]
+  const search = createDirectorySearch({ sdk, home: () => "/home/luke", base: () => "/home/luke" })
+
+  expect(await search("documents")).toEqual(["/home/luke/Documents"])
+})
+
 test("searches from an absolute root without a default base", async () => {
  const directories: string[] = []
  const sdk = {
@@ -379,7 +379,14 @@ export function createDirectorySearch(args: { sdk: ServerSDK; base: () => string
        .then((result) => result.data.map((entry) => entry.path))
        .catch(() => [])
      if (!active()) return []
-      return results.map((path) => joinPickerPath(input.directory, path)).slice(0, 50)
+      if (results.length) {
+        return results.map((path) => joinPickerPath(input.directory, path)).slice(0, 50)
+      }
+      const fallback = query
+        ? await match(input.directory, query, 50)
+        : (await directories(input.directory)).map((item) => item.absolute)
+      if (!active()) return []
+      return fallback
    }
    const segments = query.replace(/^\/+/, "").split("/")
    const head = segments.slice(0, -1).filter((part) => part && part !== ".")
@@ -103,6 +103,7 @@ describe("normalizeProviderList", () => {
    )

    expect(result.connected).toEqual(["openai"])
+    expect(result.defaultModel).toEqual({ providerID: "openai", modelID: "gpt-5" })
    expect(result.default).toEqual({ openai: "gpt-5" })
    expect(result.all.get("openai")?.models["gpt-old"]).toBeUndefined()
    expect(result.all.get("openai")?.models["gpt-5"]).toMatchObject({
@@ -113,6 +114,10 @@ describe("normalizeProviderList", () => {
      variants: { high: {} },
    })
  })
+
+  test("preserves an empty current default", () => {
+    expect(normalizeProviderList([] as ProviderListOutput["data"], [], null).defaultModel).toBeNull()
+  })
 })

 describe("directoryKey", () => {
@@ -139,6 +139,7 @@ export function normalizeProviderList(
  return {
    all,
    connected: providers.map((provider) => provider.id),
+    defaultModel: defaultModel ? { providerID: defaultModel.providerID, modelID: defaultModel.id } : null,
    default: Object.fromEntries(
      providers.flatMap((provider) => {
        const model =
@@ -6,6 +6,7 @@ import { createStore } from "solid-js/store"
 import { useModels } from "@/context/models"
 import { useSettings } from "@/context/settings"
 import { useProviders } from "@/hooks/use-providers"
+import { resolveDefaultModel } from "@/hooks/provider-catalog"
 import { Persist, persisted } from "@/utils/persist"
 import { hasCustomAgent, resolveAgent } from "./local-agent"
 import { cycleModelVariant, getConfiguredAgentVariant, resolveModelVariant } from "./model-variant"
@@ -149,10 +150,8 @@ export const { use: useLocal, provider: LocalProvider } = createSimpleContext({
    })

    const configuredModel = () => {
-      const configured = sync().data.config.model
-      if (!configured) return
-      const [providerID, modelID] = configured.split("/")
-      const model = { providerID, modelID }
+      const model = resolveDefaultModel(providers.defaultModel(), sync().data.config.model)
+      if (!model) return
      if (validModel(model)) return model
    }

@@ -1,6 +1,6 @@
 import { expect, test } from "bun:test"
 import type { NormalizedProviderListResponse } from "@opencode-ai/session-ui/context"
-import { selectProviderCatalog } from "./provider-catalog"
+import { resolveDefaultModel, selectProviderCatalog } from "./provider-catalog"

 const catalog = (id: string): NormalizedProviderListResponse => ({
  all: new Map([[id, { id, name: id, source: "api", env: [], options: {}, models: {} }]]),
@@ -57,3 +57,21 @@ test("falls back to the global catalog for route consumers", () => {
    }),
  ).toBe(global)
 })
+
+test("uses the current server default model", () => {
+  expect(resolveDefaultModel({ providerID: "openai", modelID: "gpt-5" }, "anthropic/claude")).toEqual({
+    providerID: "openai",
+    modelID: "gpt-5",
+  })
+})
+
+test("does not use legacy config when the current server has no default", () => {
+  expect(resolveDefaultModel(null, "anthropic/claude")).toBeUndefined()
+})
+
+test("uses config for legacy servers", () => {
+  expect(resolveDefaultModel(undefined, "anthropic/claude")).toEqual({
+    providerID: "anthropic",
+    modelID: "claude",
+  })
+})
@@ -25,3 +25,13 @@ export function selectProviderCatalog(input: ProviderCatalogInput) {
  if (input.explicit) return emptyProviderCatalog
  return input.global
 }
+
+export function resolveDefaultModel(
+  current: NormalizedProviderListResponse["defaultModel"],
+  legacy: string | undefined,
+) {
+  if (current !== undefined) return current ?? undefined
+  if (!legacy) return undefined
+  const [providerID, modelID] = legacy.split("/")
+  return { providerID, modelID }
+}
@@ -41,6 +41,7 @@ export function useProviders(directory: Accessor<string | undefined>) {
  return {
    all: () => providers().all,
    default: () => providers().default,
+    defaultModel: () => providers().defaultModel,
    popular: () =>
      pipe(
        providers().all,
@@ -583,7 +583,7 @@ export const dict = {
  "dialog.usageExceeded.dontShowAgain": "不再显示",

  "context.breakdown.title": "上下文细分",
-  "context.breakdown.note": "输入令牌的大致细分。“其他”包含工具定义和开销。",
+  "context.breakdown.note": "输入 token 的大致细分。“其他”包含工具定义和开销。",
  "context.breakdown.system": "系统",
  "context.breakdown.user": "用户",
  "context.breakdown.assistant": "助手",
@@ -598,18 +598,18 @@ export const dict = {
  "context.stats.provider": "提供商",
  "context.stats.model": "模型",
  "context.stats.limit": "上下文限制",
-  "context.stats.totalTokens": "令牌总数",
+  "context.stats.totalTokens": "总 token",
  "context.stats.usage": "使用率",
-  "context.stats.inputTokens": "输入令牌",
-  "context.stats.outputTokens": "输出令牌",
-  "context.stats.reasoningTokens": "推理令牌",
-  "context.stats.cacheTokens": "缓存令牌（读/写）",
+  "context.stats.inputTokens": "输入 token",
+  "context.stats.outputTokens": "输出 token",
+  "context.stats.reasoningTokens": "推理 token",
+  "context.stats.cacheTokens": "缓存 token（读/写）",
  "context.stats.userMessages": "用户消息",
  "context.stats.assistantMessages": "助手消息",
  "context.stats.totalCost": "总成本",
  "context.stats.sessionCreated": "创建时间",
  "context.stats.lastActivity": "最后活动",
-  "context.usage.tokens": "令牌数",
+  "context.usage.tokens": "Token",
  "context.usage.usage": "使用率",
  "context.usage.cost": "成本",
  "context.usage.clickToView": "点击查看上下文",
@@ -90,7 +90,20 @@ export function createHomeController() {
        const directory = directories[0]
        if (!directory) return
        const ctx = global.ensureServerCtx(conn)
-        directories.forEach((item) => ctx.projects.open(item))
+        directories.forEach((item) => {
+          if (ctx.projects.list().some((project) => project.worktree === item)) return
+          const location = { directory: item }
+          void ctx.sdk.api.file
+            .list({ path: ".", location })
+            .then(async (files) => {
+              if (files.data.length > 0) return ctx.sdk.api.project.current({ location })
+              const result = await ctx.sdk.client.project.initGit({ directory: item })
+              return result.data ?? ctx.sdk.api.project.current({ location })
+            })
+            .then((project) => ctx.sync.child(item, { bootstrap: false })[1]("project", project.id))
+            .catch(() => undefined)
+          ctx.projects.open(item)
+        })
        ctx.projects.touch(directory)
        setSelection({ server: ServerConnection.key(conn), directory })
      },
@@ -475,6 +475,10 @@ function HomeProjectRow(
      ref={sortable.ref}
      class="group/project relative flex h-7 min-w-0 items-center rounded-[6px]"
      classList={{ "z-10": sortable.isDragSource() }}
+      onContextMenu={(event) => {
+        event.preventDefault()
+        props.onSetContextMenuOpen(contextMenuID(), true)
+      }}
    >
      <HomeProjectNavButton
        type="button"
@@ -6,6 +6,7 @@ import { usePrompt } from "@/context/prompt"
 import { useSDK } from "@/context/sdk"
 import { useSync } from "@/context/sync"
 import { useProviders } from "@/hooks/use-providers"
+import { resolveDefaultModel } from "@/hooks/provider-catalog"

 export function createPromptModelSelection(input: { agent: () => { model?: ModelKey; variant?: string } | undefined }) {
  const sdk = useSDK()
@@ -21,10 +22,8 @@ export function createPromptModelSelection(input: { agent: () => { model?: Model
  }

  const configured = () => {
-    const value = sync().data.config.model
-    if (!value) return
-    const [providerID, modelID] = value.split("/")
-    const model = { providerID, modelID }
+    const model = resolveDefaultModel(providers.defaultModel(), sync().data.config.model)
+    if (!model) return
    if (valid(model)) return model
  }

@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/cli",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "bin": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/codemode",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "description": "Effect-native confined code execution over schema-described tools",
  "private": true,
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-app",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -373,7 +373,8 @@ export const dict = {
  "zen.api.error.providerNotSupported": "Provider {{provider}} not supported",
  "zen.api.error.missingApiKey": "Missing API key.",
  "zen.api.error.invalidApiKey": "Invalid API key.",
-  "zen.api.error.requestBlockedByUpstreamProvider": "Request blocked by upstream provider.",
+  "zen.api.error.requestBlockedByUpstreamProvider":
+    "This account has found to be committing fraud or is in breach of terms of services and has been blocked. If you believe this is a mistake please send an email to help@anoma.ly.",
  "zen.api.error.subscriptionQuotaExceeded": "Subscription quota exceeded. Retry in {{retryIn}}.",
  "zen.api.error.goSubscriptionRollingLimitExceeded":
    "5-hour usage limit reached. Resets in {{retryIn}}. To continue using this model now, enable usage from your available balance: {{consoleGoUrl}}",
@@ -0,0 +1,158 @@
+import type { APIEvent } from "@solidjs/start/server"
+import { and, Database, eq, isNull } from "@opencode-ai/console-core/drizzle/index.js"
+import { BillingTable, LiteTable } from "@opencode-ai/console-core/schema/billing.sql.js"
+import { KeyTable } from "@opencode-ai/console-core/schema/key.sql.js"
+import { UserTable } from "@opencode-ai/console-core/schema/user.sql.js"
+import { WorkspaceTable } from "@opencode-ai/console-core/schema/workspace.sql.js"
+import { LiteData } from "@opencode-ai/console-core/lite.js"
+import { Subscription } from "@opencode-ai/console-core/subscription.js"
+
+export async function GET(input: APIEvent) {
+  const apiKey = input.request.headers.get("authorization")?.match(/^Bearer (\S+)$/)?.[1]
+
+  if (!apiKey) {
+    return new Response(
+      JSON.stringify({
+        type: "error",
+        error: {
+          type: "AuthError",
+          message: "Missing API key.",
+        },
+      }),
+      {
+        status: 401,
+        headers: {
+          "Content-Type": "application/json",
+        },
+      },
+    )
+  }
+
+  const auth = await Database.use((tx) =>
+    tx
+      .select({
+        userID: KeyTable.userID,
+        workspaceID: KeyTable.workspaceID,
+      })
+      .from(KeyTable)
+      .innerJoin(
+        UserTable,
+        and(
+          eq(UserTable.workspaceID, KeyTable.workspaceID),
+          eq(UserTable.id, KeyTable.userID),
+          isNull(UserTable.timeDeleted),
+        ),
+      )
+      .innerJoin(WorkspaceTable, and(eq(WorkspaceTable.id, KeyTable.workspaceID), isNull(WorkspaceTable.timeDeleted)))
+      .innerJoin(
+        BillingTable,
+        and(eq(BillingTable.workspaceID, KeyTable.workspaceID), isNull(BillingTable.timeDeleted)),
+      )
+      .where(and(eq(KeyTable.key, apiKey), isNull(KeyTable.timeDeleted)))
+      .then((rows) => rows[0]),
+  )
+
+  if (!auth) {
+    return new Response(
+      JSON.stringify({
+        type: "error",
+        error: {
+          type: "AuthError",
+          message: "Unauthorized",
+        },
+      }),
+      {
+        status: 401,
+        headers: {
+          "Content-Type": "application/json",
+        },
+      },
+    )
+  }
+
+  const row = await Database.use((tx) =>
+    tx
+      .select({
+        timeCreated: LiteTable.timeCreated,
+        rollingUsage: LiteTable.rollingUsage,
+        weeklyUsage: LiteTable.weeklyUsage,
+        monthlyUsage: LiteTable.monthlyUsage,
+        timeRollingUpdated: LiteTable.timeRollingUpdated,
+        timeWeeklyUpdated: LiteTable.timeWeeklyUpdated,
+        timeMonthlyUpdated: LiteTable.timeMonthlyUpdated,
+      })
+      .from(LiteTable)
+      .where(
+        and(
+          eq(LiteTable.workspaceID, auth.workspaceID),
+          eq(LiteTable.userID, auth.userID),
+          isNull(LiteTable.timeDeleted),
+        ),
+      )
+      .then((rows) => rows[0]),
+  )
+
+  if (!row) {
+    return new Response(
+      JSON.stringify({
+        type: "error",
+        error: {
+          type: "EntitlementError",
+          message: "OpenCode Go subscription required.",
+        },
+      }),
+      {
+        status: 403,
+        headers: {
+          "Content-Type": "application/json",
+        },
+      },
+    )
+  }
+
+  const limits = LiteData.getLimits()
+
+  return new Response(
+    JSON.stringify({
+      usage: {
+        rolling: formatUsage(
+          Subscription.analyzeRollingUsage({
+            limit: limits.rollingLimit,
+            window: limits.rollingWindow,
+            usage: row.rollingUsage ?? 0,
+            timeUpdated: row.timeRollingUpdated ?? new Date(),
+          }),
+        ),
+        weekly: formatUsage(
+          Subscription.analyzeWeeklyUsage({
+            limit: limits.weeklyLimit,
+            usage: row.weeklyUsage ?? 0,
+            timeUpdated: row.timeWeeklyUpdated ?? new Date(),
+          }),
+        ),
+        monthly: formatUsage(
+          Subscription.analyzeMonthlyUsage({
+            limit: limits.monthlyLimit,
+            usage: row.monthlyUsage ?? 0,
+            timeUpdated: row.timeMonthlyUpdated ?? new Date(),
+            timeSubscribed: row.timeCreated,
+          }),
+        ),
+      },
+    }),
+    {
+      status: 200,
+      headers: {
+        "Content-Type": "application/json",
+      },
+    },
+  )
+}
+
+function formatUsage(usage: { status: "ok" | "rate-limited"; resetInSec: number; usagePercent: number }) {
+  return {
+    status: usage.status,
+    percent: usage.usagePercent,
+    resetsAt: new Date(Date.now() + usage.resetInSec * 1000).toISOString(),
+  }
+}
@@ -139,7 +139,7 @@ export async function handler(
    if (
      authInfo &&
      opts.modelList === "lite" &&
-      modelInfo.id === "deepseek-v4-flash" &&
+      ["deepseek-v4-flash", "deepseek-v4-pro"].includes(modelInfo.id) &&
      !allowedRegions?.includes("cn")
    )
      throw new RegionError(
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/console-core",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-function",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "$schema": "https://json.schemastore.org/package.json",
  "private": true,
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-mail",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "dependencies": {
    "@jsx-email/all": "2.2.3",
    "@jsx-email/cli": "1.4.3",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/console-support",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "name": "@opencode-ai/core",
  "type": "module",
  "license": "MIT",
@@ -30,15 +30,11 @@ Guidelines:

 Complete the user's search request efficiently and report your findings clearly.`

-const PROMPT_COMPACTION = `You are an anchored context summarization assistant for coding sessions.
-
-Summarize only the conversation history you are given. The newest turns may be kept verbatim outside your summary, so focus on the older context that still matters for continuing the work.
-
-If the prompt includes a <previous-summary> block, treat it as the current anchored summary. Update it with the new history by preserving still-true details, removing stale details, and merging in new facts.
+const PROMPT_COMPACTION = `You are a context summarization agent. You are given a conversation between a user and an agent. Your goal is to produce a structured summary matching the format specified so another coding agent can continue the work.

 Always follow the exact output structure requested by the user prompt. Keep every section, preserve exact file paths and identifiers when known, and prefer terse bullets over paragraphs.

-Do not answer the conversation itself. Do not mention that you are summarizing, compacting, or merging context. Respond in the same language as the conversation.`
+Do not continue the conversation. Do not respond to any questions in the conversation. Only output the structured summary in the exact format requested by the user prompt. Respond in the same language as the conversation.`

 const PROMPT_TITLE = `You are a title generator. You output ONLY a thread title. Nothing else.

@@ -40,7 +40,7 @@ already-loaded config until then.
 | Scope                         | Path                                                                                                                      |
 | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------- |
 | Project config                | `./opencode.json`, `./opencode.jsonc`, or `.opencode/opencode.json` (opencode walks up from the cwd to the worktree root) |
-| Global config                 | `~/.config/opencode/opencode.json` (NOT `~/.opencode/`)                                                                   |
+| Global config                 | `~/.config/opencode/opencode.json` or `~/.config/opencode/opencode.jsonc` (NOT `~/.opencode/`)                            |
 | Project agents                | `.opencode/agent/<name>.md` or `.opencode/agents/<name>.md`                                                               |
 | Global agents                 | `~/.config/opencode/agent(s)/<name>.md`                                                                                   |
 | Project commands              | `.opencode/command/<name>.md` or `.opencode/commands/<name>.md`                                                           |
@@ -264,7 +264,10 @@ const layer = Layer.effect(
                }),
                line: match.line_number,
                offset: match.absolute_offset,
-                text: match.lines.text.length > 2_000 ? match.lines.text.slice(0, 2_000) + "..." : match.lines.text,
+                text:
+                  match.lines.text.length > 2_000
+                    ? match.lines.text.slice(0, 2_000).replace(/[\uD800-\uDBFF]$/, "") + "..."
+                    : match.lines.text,
                submatches: match.submatches.map((submatch) => ({
                  text: submatch.match.text,
                  start: submatch.start,
@@ -44,6 +44,15 @@ Rules:
 - Use terse bullets, not prose paragraphs.
 - Preserve exact file paths, symbols, commands, error strings, URLs, and identifiers when known.
 - Do not mention the summary process or that context was compacted.`
+const SUMMARY_UPDATE_INSTRUCTIONS = `The <prior-summary> summarizes everything that happened before the <conversation>. Construct a new summary that combines both. The <prior-summary> is discarded after this: anything you do not carry into the new summary is lost.
+
+When combining:
+- Carry forward objectives, constraints, user directives, decisions, and parallel workstreams from the <prior-summary> even when the <conversation> does not mention them. Drop only what is finished and no longer needed.
+- The <conversation> is more recent than the <prior-summary>. Where they conflict, the conversation wins: state the corrected fact and drop the old claim.
+- Add new progress, decisions, constraints, and context from the conversation.
+- Move completed work from "Active" to "Completed".
+- If a blocker has been resolved, update the summary to reflect that while keeping any details still needed to continue the work.
+- Update "Objective" and "Next Move" to reflect the current work state.`

 type Entry = {
  readonly seq: number
@@ -136,36 +145,45 @@ const select = (
  if (conversation.length === 0) return
  let total = 0
  let split = conversation.length
-  let splitPrefix = ""
-  let splitSuffix = ""
  for (let index = conversation.length - 1; index >= 0; index--) {
    const next = total + Token.estimate(conversation[index])
-    if (next > tokens) {
-      const remaining = Math.max(0, tokens - total) * 4
-      if (remaining > 0) {
-        splitPrefix = conversation[index].slice(0, -remaining)
-        splitSuffix = conversation[index].slice(-remaining)
-        split = index + 1
-      }
-      break
-    }
+    if (next > tokens) break
    total = next
    split = index
  }
  return {
-    head: [...conversation.slice(0, split), splitPrefix].filter(Boolean).join("\n\n"),
-    recent: [splitSuffix, ...conversation.slice(split)].filter(Boolean).join("\n\n"),
+    head: conversation.slice(0, split).join("\n\n"),
+    recent: conversation.slice(split).join("\n\n"),
  }
 }

-export const buildPrompt = (input: { readonly previousSummary?: string; readonly context: readonly string[] }) =>
-  [
-    input.previousSummary
-      ? `Update the anchored summary below using the conversation history above.\nPreserve still-true details, remove stale details, and merge in the new facts.\n<previous-summary>\n${input.previousSummary}\n</previous-summary>`
-      : "Create a new anchored summary from the conversation history.",
+export const buildPrompt = (input: { readonly previousSummary?: string; readonly context: readonly string[] }) => {
+  const conversation = `Here is the conversation so far:\n\n<conversation>\n${input.context.join("\n\n")}\n</conversation>`
+  if (!input.previousSummary)
+    return [
+      conversation,
+      "Create a new anchored summary from the conversation history in the <conversation> tags above so another coding agent can continue the work.",
+      SUMMARY_TEMPLATE,
+    ].join("\n\n")
+  return [
+    conversation,
+    `Here is the summary of the conversation before the <conversation> above:\n\n<prior-summary>\n${input.previousSummary}\n</prior-summary>`,
+    SUMMARY_UPDATE_INSTRUCTIONS,
    SUMMARY_TEMPLATE,
-    ...input.context,
  ].join("\n\n")
+}
+
+export const buildReplayPrompt = (previousSummary?: string) => {
+  const instruction =
+    "Create a new anchored summary from the conversation messages above so another coding agent can continue the work."
+  if (!previousSummary) return [instruction, SUMMARY_TEMPLATE].join("\n\n")
+  return [
+    instruction,
+    `Here is the summary of the conversation before the messages above:\n\n<prior-summary>\n${previousSummary}\n</prior-summary>`,
+    SUMMARY_UPDATE_INSTRUCTIONS,
+    SUMMARY_TEMPLATE,
+  ].join("\n\n")
+}

 export const make = (dependencies: Dependencies) => {
  const config = settings(dependencies.config)
@@ -156,7 +156,7 @@ export const Info = Schema.Struct({
      }),
      tail_turns: Schema.optional(NonNegativeInt).annotate({
        description:
-          "Number of recent user turns, including their following assistant/tool responses, to keep verbatim during compaction (default: 2)",
+          "Maximum number of recent user turns, including their following assistant/tool responses, to keep verbatim during compaction. By default retention is limited only by the preserved token budget.",
      }),
      preserve_recent_tokens: Schema.optional(NonNegativeInt).annotate({
        description: "Maximum number of tokens from recent turns to preserve verbatim after compaction",
@@ -0,0 +1,28 @@
+import { createGroq } from "@ai-sdk/groq"
+import { expect, test } from "bun:test"
+
+test("Groq passes through unknown reasoning effort", async () => {
+  let body: Record<string, unknown> | undefined
+  const mockFetch = Object.assign(
+    async (_input: Parameters<typeof fetch>[0], init?: RequestInit) => {
+      body = JSON.parse(String(init?.body))
+      return Response.json({
+        id: "response-1",
+        created: 0,
+        model: "openai/gpt-oss-120b",
+        object: "chat.completion",
+        choices: [{ index: 0, message: { role: "assistant", content: "Hello" }, finish_reason: "stop" }],
+        usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 },
+      })
+    },
+    { preconnect: fetch.preconnect },
+  )
+  const model = createGroq({ apiKey: "test", fetch: mockFetch })("openai/gpt-oss-120b")
+
+  await model.doGenerate({
+    prompt: [{ role: "user", content: [{ type: "text", text: "Hello" }] }],
+    providerOptions: { groq: { reasoningEffort: "custom" } },
+  })
+
+  expect(body?.reasoning_effort).toBe("custom")
+})
@@ -27,6 +27,32 @@ test("Mistral sends promptCacheKey as prompt_cache_key", async () => {
  expect(body?.prompt_cache_key).toBe("session-123")
 })

+test("Mistral passes through unknown reasoning effort", async () => {
+  let body: Record<string, unknown> | undefined
+  const mockFetch = Object.assign(
+    async (_input: Parameters<typeof fetch>[0], init?: RequestInit) => {
+      body = JSON.parse(String(init?.body))
+      return Response.json({
+        id: "response-1",
+        created: 0,
+        model: "mistral-large-latest",
+        object: "chat.completion",
+        choices: [{ index: 0, message: { role: "assistant", content: "Hello" }, finish_reason: "stop" }],
+        usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 },
+      })
+    },
+    { preconnect: fetch.preconnect },
+  )
+  const model = createMistral({ apiKey: "test", fetch: mockFetch })("mistral-large-latest")
+
+  await model.doGenerate({
+    prompt: [{ role: "user", content: [{ type: "text", text: "Hello" }] }],
+    providerOptions: { mistral: { reasoningEffort: "custom" } },
+  })
+
+  expect(body?.reasoning_effort).toBe("custom")
+})
+
 test("Mistral round-trips native reasoning in assistant history", async () => {
  let body: { messages?: unknown[] } | undefined
  const mockFetch = Object.assign(
@@ -30,3 +30,56 @@ test("xAI Responses sends promptCacheKey as prompt_cache_key", async () => {

  expect(body?.prompt_cache_key).toBe("session-123")
 })
+
+test("xAI Responses passes through xhigh reasoning effort", async () => {
+  let body: Record<string, unknown> | undefined
+  const mockFetch = Object.assign(
+    async (_input: Parameters<typeof fetch>[0], init?: RequestInit) => {
+      body = JSON.parse(String(init?.body))
+      return Response.json({
+        id: "response-1",
+        created_at: 0,
+        model: "grok-4",
+        object: "response",
+        output: [],
+        usage: { input_tokens: 1, output_tokens: 0 },
+        status: "completed",
+      })
+    },
+    { preconnect: fetch.preconnect },
+  )
+  const model = createXai({ apiKey: "test", fetch: mockFetch }).responses("grok-4")
+
+  await model.doGenerate({
+    prompt: [{ role: "user", content: [{ type: "text", text: "Hello" }] }],
+    providerOptions: { xai: { reasoningEffort: "xhigh" } },
+  })
+
+  expect(body?.reasoning).toEqual({ effort: "xhigh" })
+})
+
+test("xAI Chat passes through xhigh reasoning effort", async () => {
+  let body: Record<string, unknown> | undefined
+  const mockFetch = Object.assign(
+    async (_input: Parameters<typeof fetch>[0], init?: RequestInit) => {
+      body = JSON.parse(String(init?.body))
+      return Response.json({
+        id: "chat-1",
+        created: 0,
+        model: "grok-4",
+        object: "chat.completion",
+        choices: [{ index: 0, message: { role: "assistant", content: "Hello" }, finish_reason: "stop" }],
+        usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 },
+      })
+    },
+    { preconnect: fetch.preconnect },
+  )
+  const model = createXai({ apiKey: "test", fetch: mockFetch }).chat("grok-4")
+
+  await model.doGenerate({
+    prompt: [{ role: "user", content: [{ type: "text", text: "Hello" }] }],
+    providerOptions: { xai: { reasoningEffort: "xhigh" } },
+  })
+
+  expect(body?.reasoning_effort).toBe("xhigh")
+})
@@ -62,4 +62,24 @@ describe("Ripgrep", () => {
      (tmp) => Effect.promise(() => tmp[Symbol.asyncDispose]()),
    ),
  )
+  it.live("does not split surrogate pairs in oversized line previews", () =>
+    Effect.acquireUseRelease(
+      Effect.promise(() => tmpdir()),
+      (tmp) =>
+        Effect.gen(function* () {
+          yield* Effect.promise(() =>
+            fs.writeFile(path.join(tmp.path, "unicode.txt"), `needle${"x".repeat(1_993)}😀\n`),
+          )
+
+          const matches = yield* (yield* Ripgrep.Service).grep({
+            cwd: tmp.path,
+            pattern: "needle",
+            limit: 10,
+          })
+
+          expect(matches[0]?.text).toBe(`needle${"x".repeat(1_993)}...`)
+        }),
+      (tmp) => Effect.promise(() => tmp[Symbol.asyncDispose]()),
+    ),
+  )
 })
@@ -4,12 +4,32 @@ import { SessionCompaction } from "@opencode-ai/core/session/compaction"
 test("compaction prompt preserves detailed work state and relevant files", () => {
  const prompt = SessionCompaction.buildPrompt({ context: ["conversation history"] })

+  expect(prompt).toStartWith(
+    "Here is the conversation so far:\n\n<conversation>\nconversation history\n</conversation>",
+  )
+  expect(prompt.indexOf("</conversation>")).toBeLessThan(prompt.indexOf("Create a new anchored summary"))
+  expect(prompt).toContain("conversation history in the <conversation> tags above")
  expect(prompt).toContain("## Work State\n### Completed")
  expect(prompt).toContain("### Active")
  expect(prompt).toContain("### Blocked")
  expect(prompt).toContain("## Relevant Files")
 })

+test("compaction prompt gives update instructions for a prior summary", () => {
+  const prompt = SessionCompaction.buildPrompt({
+    context: ["new conversation"],
+    previousSummary: "existing summary",
+  })
+
+  expect(prompt.indexOf("<conversation>")).toBeLessThan(prompt.indexOf("<prior-summary>"))
+  expect(prompt.indexOf("</prior-summary>")).toBeLessThan(prompt.indexOf("The <prior-summary> summarizes"))
+  expect(prompt).toContain(
+    "Carry forward objectives, constraints, user directives, decisions, and parallel workstreams from the <prior-summary>",
+  )
+  expect(prompt).toContain('Move completed work from "Active" to "Completed".')
+  expect(prompt).toContain('Update "Objective" and "Next Move" to reflect the current work state.')
+})
+
 test("compaction describes tool media without embedding base64", () => {
  const base64 = "iVBORw0KGgoAAAANSUhEUgAAAAEAAAAB"
  const serialized = SessionCompaction.serializeToolContent([
@@ -1135,7 +1135,7 @@ describe("SessionRunnerLLM", () => {

      expect(requests).toHaveLength(2)
      expect(userTexts(requests[0])[0]).toContain(
-        "<previous-summary>\n## Objective\n- Preserve the task\n</previous-summary>",
+        "<prior-summary>\n## Objective\n- Preserve the task\n</prior-summary>",
      )
      expect(userTexts(requests[0])[0]).toContain("Recent exact request")
      expect((yield* (yield* SessionStore.Service).context(sessionID))[0]).toMatchObject({
@@ -1145,6 +1145,67 @@ describe("SessionRunnerLLM", () => {
    }),
  )

+  it.effect("retains only complete serialized messages during compaction", () =>
+    Effect.gen(function* () {
+      yield* setup
+      const session = yield* SessionV2.Service
+      const earlier = `EARLIER_BOUNDARY ${"a".repeat(3_000)} EARLIER_END`
+      const recent = `RECENT_BOUNDARY ${"b".repeat(3_000)} RECENT_END`
+      response = fragmentFixture("text", "text-earlier", ["Earlier answer"]).completeEvents
+      yield* session.prompt({ sessionID, prompt: Prompt.make({ text: earlier }), resume: false })
+      yield* session.resume(sessionID)
+
+      currentModel = compactModel
+      requests.length = 0
+      responses = [
+        fragmentFixture("text", "text-summary", ["## Objective\n- Preserve the task"]).completeEvents,
+        fragmentFixture("text", "text-final", ["Continued"]).completeEvents,
+      ]
+      yield* session.prompt({ sessionID, prompt: Prompt.make({ text: recent }), resume: false })
+      yield* session.resume(sessionID)
+
+      expect(requests).toHaveLength(2)
+      const summary = userTexts(requests[0])[0]
+      const continuation = userTexts(requests[1])[0]
+      expect(summary.match(/EARLIER_BOUNDARY/g)).toHaveLength(1)
+      expect(summary).toContain(`EARLIER_BOUNDARY ${"a".repeat(3_000)} EARLIER_END`)
+      expect(summary).not.toContain("RECENT_BOUNDARY")
+      expect(continuation).not.toContain("EARLIER_BOUNDARY")
+      expect(continuation).not.toContain("EARLIER_END")
+      expect(continuation).toContain("<recent-context>\n[Assistant]: Earlier answer")
+      expect(continuation).toContain(`RECENT_BOUNDARY ${"b".repeat(3_000)} RECENT_END`)
+    }),
+  )
+
+  it.effect("summarizes an oversized newest message without retaining a fragment", () =>
+    Effect.gen(function* () {
+      yield* setup
+      const session = yield* SessionV2.Service
+      response = fragmentFixture("text", "text-earlier", ["Earlier answer"]).completeEvents
+      yield* session.prompt({ sessionID, prompt: Prompt.make({ text: "Earlier question" }), resume: false })
+      yield* session.resume(sessionID)
+
+      const oversized = `OVERSIZED_BOUNDARY ${"x".repeat(4_500)} OVERSIZED_END`
+      currentModel = compactModel
+      requests.length = 0
+      responses = [
+        fragmentFixture("text", "text-summary", ["## Objective\n- Preserve the task"]).completeEvents,
+        fragmentFixture("text", "text-final", ["Continued"]).completeEvents,
+      ]
+      yield* session.prompt({ sessionID, prompt: Prompt.make({ text: oversized }), resume: false })
+      yield* session.resume(sessionID)
+
+      expect(requests).toHaveLength(2)
+      const summary = userTexts(requests[0])[0]
+      const continuation = userTexts(requests[1])[0]
+      expect(summary.match(/OVERSIZED_BOUNDARY/g)).toHaveLength(1)
+      expect(summary).toContain(oversized)
+      expect(continuation).not.toContain("OVERSIZED_BOUNDARY")
+      expect(continuation).not.toContain("OVERSIZED_END")
+      expect(continuation).toContain("<recent-context>\n\n</recent-context>")
+    }),
+  )
+
  it.effect("forces one compaction and retries after provider context overflow", () =>
    Effect.gen(function* () {
      const session = yield* setupOverflowRecovery
@@ -1,7 +1,7 @@
 {
  "name": "@opencode-ai/desktop",
  "private": true,
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "homepage": "https://opencode.ai",
@@ -6,7 +6,7 @@ import { homedir, tmpdir } from "node:os"
 import { join } from "node:path"
 import { getCACertificates, setDefaultCACertificates } from "node:tls"
 import type { Event } from "electron"
-import { app } from "electron"
+import { app, BrowserWindow } from "electron"

 import { Deferred, Effect, Fiber } from "effect"
 import contextMenu from "electron-context-menu"
@@ -408,6 +408,15 @@ const main = Effect.gen(function* () {

  yield* Fiber.await(loadingTask)

+  app.on("window-all-closed", () => {
+    if (process.platform === "darwin") return
+    app.quit()
+  })
+  app.on("activate", () => {
+    if (BrowserWindow.getAllWindows().length > 0) return
+    restoreMainWindows()
+  })
+
  const windows = restoreMainWindows()
  if (windows.length) createMenu(menuDeps)
 })
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "name": "@opencode-ai/effect-drizzle-sqlite",
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "name": "@opencode-ai/effect-sqlite-node",
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/enterprise",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/function",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "$schema": "https://json.schemastore.org/package.json",
  "private": true,
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "name": "@opencode-ai/http-recorder",
  "description": "Record and replay Effect HTTP client traffic with deterministic cassettes",
  "type": "module",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "name": "@opencode-ai/llm",
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "name": "opencode",
  "type": "module",
  "license": "MIT",
@@ -216,21 +216,6 @@ const layer = Layer.effect(
            mode: "subagent",
            native: true,
          },
-          compaction: {
-            name: "compaction",
-            mode: "primary",
-            native: true,
-            hidden: true,
-            prompt: PROMPT_COMPACTION,
-            permission: Permission.merge(
-              defaults,
-              Permission.fromConfig({
-                "*": "deny",
-              }),
-              user,
-            ),
-            options: {},
-          },
          title: {
            name: "title",
            mode: "primary",
@@ -270,6 +255,26 @@ const layer = Layer.effect(
            continue
          }
          let item = agents[key]
+          // The compaction agent is not registered by default: compaction inherits
+          // the session's active agent so its request stays prefix-identical for
+          // prompt caching. An explicit `agent.compaction` config opts into the
+          // legacy dedicated agent, seeded here so overrides merge as before.
+          if (!item && key === "compaction")
+            item = agents[key] = {
+              name: "compaction",
+              mode: "primary",
+              native: true,
+              hidden: true,
+              prompt: PROMPT_COMPACTION,
+              permission: Permission.merge(
+                defaults,
+                Permission.fromConfig({
+                  "*": "deny",
+                }),
+                user,
+              ),
+              options: {},
+            }
          if (!item)
            item = agents[key] = {
              name: key,
@@ -1,9 +1,5 @@
-You are an anchored context summarization assistant for coding sessions.
-
-Summarize only the conversation history you are given. The newest turns may be kept verbatim outside your summary, so focus on the older context that still matters for continuing the work.
-
-If the prompt includes a <previous-summary> block, treat it as the current anchored summary. Update it with the new history by preserving still-true details, removing stale details, and merging in new facts.
+You are a context summarization agent. You are given a conversation between a user and an agent. Your goal is to produce a structured summary matching the format specified so another coding agent can continue the work.

 Always follow the exact output structure requested by the user prompt. Keep every section, preserve exact file paths and identifiers when known, and prefer terse bullets over paragraphs.

-Do not answer the conversation itself. Do not mention that you are summarizing, compacting, or merging context. Respond in the same language as the conversation.
+Do not continue the conversation. Do not respond to any questions in the conversation. Only output the structured summary in the exact format requested by the user prompt. Respond in the same language as the conversation.
@@ -37,22 +37,11 @@ export function schema<S extends EffectSchema.Decoder<unknown, never>>(
  data: unknown,
  source: string,
 ): DeepMutable<S["Type"]> {
-  const extra = topLevelExtraKeys(schema, data)
-  if (extra.length) {
-    throw new InvalidError({
-      path: source,
-      issues: [
-        {
-          code: "unrecognized_keys",
-          keys: extra,
-          path: [],
-          message: `Unrecognized key${extra.length === 1 ? "" : "s"}: ${extra.join(", ")}`,
-        },
-      ],
-    })
-  }
-
-  const decoded = EffectSchema.decodeUnknownExit(schema)(data, { errors: "all", propertyOrder: "original" })
+  const decoded = EffectSchema.decodeUnknownExit(schema)(data, {
+    errors: "all",
+    onExcessProperty: "ignore",
+    propertyOrder: "original",
+  })
  if (Exit.isSuccess(decoded)) return decoded.value as DeepMutable<S["Type"]>
  const error = Cause.squash(decoded.cause)

@@ -70,10 +59,3 @@ export function schema<S extends EffectSchema.Decoder<unknown, never>>(
    { cause: error },
  )
 }
-
-function topLevelExtraKeys(schema: EffectSchema.Top, data: unknown) {
-  if (typeof data !== "object" || data === null || Array.isArray(data)) return []
-  if (schema.ast._tag !== "Objects" || schema.ast.indexSignatures.length > 0) return []
-  const known = new Set(schema.ast.propertySignatures.map((item) => String(item.name)))
-  return Object.keys(data).filter((key) => !known.has(key))
-}
@@ -88,6 +88,9 @@ function build(key: string, remote: SelectableItem, url: string, prev?: Model):
  const image =
    (remote.capabilities.supports.vision ?? false) ||
    (remote.capabilities.limits.vision?.supported_media_types ?? []).some((item) => item.startsWith("image/"))
+  const pdf =
+    (remote.capabilities.supports.vision ?? false) &&
+    (remote.capabilities.limits.vision?.supported_media_types?.includes("application/pdf") ?? false)

  const isMsgApi = remote.supported_endpoints?.includes("/v1/messages")
  const endpoint: CopilotEndpoint | undefined = isMsgApi
@@ -127,7 +130,7 @@ function build(key: string, remote: SelectableItem, url: string, prev?: Model):
        audio: false,
        image,
        video: false,
-        pdf: false,
+        pdf,
      },
      output: {
        text: true,
@@ -84,6 +84,8 @@ function sdkKey(npm: string): string | undefined {
      return "gateway"
    case "@openrouter/ai-sdk-provider":
      return "openrouter"
+    case "merge-gateway-ai-sdk-provider":
+      return "mergeGateway"
    case "ai-gateway-provider":
      // ai-gateway-provider/unified wraps createOpenAICompatible({ name: "Unified" }),
      // and @ai-sdk/openai-compatible parses compatibleOptions from one of
@@ -551,6 +553,12 @@ export function topP(model: Provider.Model) {
  if (["minimax-m2", "kimi-k2.5", "kimi-k2p5", "kimi-k2-5"].some((s) => id.includes(s))) {
    return 0.95
  }
+  if (
+    ["deepseek-v4-flash-0731", "deepseek-v4-flash:0731"].some((name) => id.includes(name)) ||
+    (id.includes("deepseek-v4-flash") && (model.providerID === "deepseek" || model.providerID.startsWith("opencode")))
+  ) {
+    return 0.95
+  }
  return undefined
 }

@@ -1766,6 +1774,7 @@ function reasoningEffort(model: Provider.Model, effort: string) {
    case "@ai-sdk/togetherai":
    case "venice-ai-sdk-provider":
    case "ai-gateway-provider":
+    case "merge-gateway-ai-sdk-provider":
      return { reasoningEffort: effort }
    case "@ai-sdk/cohere":
    case "@ai-sdk/perplexity":
@@ -20,7 +20,7 @@ import { RuntimeFlags } from "@/effect/runtime-flags"
 import { EventV2Bridge } from "@/event-v2-bridge"
 import { ProviderV2 } from "@opencode-ai/core/provider"
 import { ModelV2 } from "@opencode-ai/core/model"
-import { buildPrompt } from "@opencode-ai/core/session/compaction"
+import { buildPrompt, buildReplayPrompt } from "@opencode-ai/core/session/compaction"
 import { SessionCompactionEvent } from "@opencode-ai/schema/session-compaction-event"

 export const Event = SessionCompactionEvent
@@ -29,9 +29,8 @@ export const PRUNE_MINIMUM = 20_000
 export const PRUNE_PROTECT = 40_000
 const TOOL_OUTPUT_MAX_CHARS = 2_000
 const PRUNE_PROTECTED_TOOLS = ["skill"]
-const DEFAULT_TAIL_TURNS = 2
 const MIN_PRESERVE_RECENT_TOKENS = 2_000
-const MAX_PRESERVE_RECENT_TOKENS = 8_000
+const MAX_PRESERVE_RECENT_TOKENS = 15_000
 type Turn = {
  start: number
  end: number
@@ -163,6 +162,15 @@ function splitTurn(input: {
  })
 }

+export type Run = (input: {
+  user: SessionV1.User
+  agent: Agent.Info
+  model: Provider.Model
+  processor: SessionProcessor.Handle
+  messages: SessionV1.WithParts[]
+  prompt: string
+}) => Effect.Effect<SessionProcessor.Result>
+
 export interface Interface {
  readonly isOverflow: (input: {
    tokens: SessionV1.Assistant["tokens"]
@@ -175,6 +183,7 @@ export interface Interface {
    sessionID: SessionID
    auto: boolean
    overflow?: boolean
+    run?: Run
  }) => Effect.Effect<"continue" | "stop">
  readonly create: (input: {
    sessionID: SessionID
@@ -226,27 +235,22 @@ const layer = Layer.effect(
      cfg: ConfigV1.Info
      model: Provider.Model
    }) {
-      const limit = input.cfg.compaction?.tail_turns ?? DEFAULT_TAIL_TURNS
-      if (limit <= 0) return { head: input.messages, tail_start_id: undefined }
+      const limit = input.cfg.compaction?.tail_turns
+      if (limit !== undefined && limit <= 0) return { head: input.messages, tail_start_id: undefined }
      const budget = preserveRecentBudget({ cfg: input.cfg, model: input.model })
      const all = turns(input.messages)
      if (!all.length) return { head: input.messages, tail_start_id: undefined }
-      const recent = all.slice(-limit)
-      const sizes = yield* Effect.forEach(
-        recent,
-        (turn) =>
-          estimate({
-            messages: input.messages.slice(turn.start, turn.end),
-            model: input.model,
-          }),
-        { concurrency: 1 },
-      )
+      const recent = limit === undefined ? all : all.slice(-limit)

      let total = 0
      let keep: Tail | undefined
      for (let i = recent.length - 1; i >= 0; i--) {
        const turn = recent[i]!
-        const size = sizes[i]
+        // estimate lazily so cost stays proportional to the retained tail, not the whole session
+        const size = yield* estimate({
+          messages: input.messages.slice(turn.start, turn.end),
+          model: input.model,
+        })
        if (total + size <= budget) {
          total += size
          keep = { start: turn.start, id: turn.id }
@@ -328,6 +332,7 @@ const layer = Layer.effect(
      sessionID: SessionID
      auto: boolean
      overflow?: boolean
+      run?: Run
    }) {
      const parent = input.messages.findLast((m) => m.info.id === input.parentID)
      if (!parent || parent.info.role !== "user") {
@@ -361,11 +366,13 @@ const layer = Layer.effect(
        }
      }

-      const agent = yield* agents.get("compaction")
-      const model = agent.model
-        ? yield* provider.getModel(agent.model.providerID, agent.model.modelID).pipe(Effect.orDie)
-        : yield* provider.getModel(userMessage.model.providerID, userMessage.model.modelID).pipe(Effect.orDie)
      const cfg = yield* config.get()
+      const configured = cfg.agent?.compaction !== undefined
+      const agent = yield* agents.get(configured ? "compaction" : userMessage.agent)
+      const model =
+        configured && agent.model
+          ? yield* provider.getModel(agent.model.providerID, agent.model.modelID).pipe(Effect.orDie)
+          : yield* provider.getModel(userMessage.model.providerID, userMessage.model.modelID).pipe(Effect.orDie)
      const history = compactionPart && messages.at(-1)?.info.id === input.parentID ? messages.slice(0, -1) : messages
      const prior = completedCompactions(history)
      const hidden = new Set(prior.flatMap((item) => [item.userIndex, item.assistantIndex]))
@@ -381,10 +388,18 @@ const layer = Layer.effect(
        { sessionID: input.sessionID },
        { context: [], prompt: undefined },
      )
-      const nextPrompt = compacting.prompt ?? buildPrompt({ previousSummary, context: compacting.context })
      const msgs = structuredClone(selected.head)
      yield* plugin.trigger("experimental.chat.messages.transform", {}, { messages: msgs })
-      const conversation = msgs.map(serialize).filter(Boolean).join("\n\n")
+      // A configured compaction agent keeps the legacy serialized-transcript path.
+      const run = configured ? undefined : input.run
+      const conversation = run ? "" : msgs.map(serialize).filter(Boolean).join("\n\n")
+      const nextPrompt = [
+        compacting.prompt ??
+          (run ? buildReplayPrompt(previousSummary) : buildPrompt({ previousSummary, context: [conversation] })),
+        ...compacting.context,
+      ]
+        .filter(Boolean)
+        .join("\n\n")
      const ctx = yield* InstanceState.context
      const msg: SessionV1.Assistant = {
        id: MessageID.ascending(),
@@ -418,27 +433,43 @@ const layer = Layer.effect(
        sessionID: input.sessionID,
        model,
      })
-      const result = yield* processor.process({
-        user: userMessage,
-        agent,
-        sessionID: input.sessionID,
-        tools: {},
-        system: [],
-        messages: [
-          {
-            role: "user",
-            content: [
+      // The synthetic compaction marker carries no per-turn system/tools settings.
+      const active = msgs.findLast(
+        (message) => message.info.role === "user" && !message.parts.some((part) => part.type === "compaction"),
+      )
+      const result = run
+        ? yield* run({
+            user: active?.info.role === "user" ? active.info : userMessage,
+            agent,
+            model,
+            processor,
+            messages: msgs,
+            prompt: nextPrompt,
+          })
+        : yield* processor.process({
+            user: userMessage,
+            agent,
+            sessionID: input.sessionID,
+            tools: {},
+            system: [],
+            messages: [
              {
-                type: "text",
-                text: [nextPrompt, "The following is the conversation history:", conversation]
-                  .filter(Boolean)
-                  .join("\n\n"),
+                role: "user",
+                content: [
+                  {
+                    type: "text",
+                    text: [
+                      nextPrompt,
+                      ...(compacting.prompt ? ["The following is the conversation history:", conversation] : []),
+                    ]
+                      .filter(Boolean)
+                      .join("\n\n"),
+                  },
+                ],
              },
            ],
-          },
-        ],
-        model,
-      })
+            model,
+          })

      if (result === "compact") {
        processor.message.error = new SessionV1.ContextOverflowError({
@@ -149,6 +149,45 @@ const layer = Layer.effect(
      } satisfies TaskPromptOps
    })

+    // One assembly path keeps compaction requests prefix-identical to normal turns for prompt caching.
+    const prepare = Effect.fn("SessionPrompt.prepare")(function* (input: {
+      session: Session.Info
+      agent: Agent.Info
+      model: Provider.Model
+      processor: SessionProcessor.Handle
+      messages: SessionV1.WithParts[]
+    }) {
+      const lastUserMsg = input.messages.findLast((message) => message.info.role === "user")
+      const tools = yield* SessionTools.resolve({
+        agent: input.agent,
+        session: input.session,
+        model: input.model,
+        processor: input.processor,
+        bypassAgentCheck: lastUserMsg?.parts.some((part) => part.type === "agent") ?? false,
+        messages: input.messages,
+        promptOps: yield* ops(),
+      }).pipe(
+        Effect.provideService(Plugin.Service, plugin),
+        Effect.provideService(Permission.Service, permission),
+        Effect.provideService(ToolRegistry.Service, registry),
+        Effect.provideService(MCP.Service, mcp),
+        Effect.provideService(Truncate.Service, truncate),
+        Effect.provideService(RuntimeFlags.Service, flags),
+      )
+      const [skills, env, instructions, mcpInstructions, modelMessages] = yield* Effect.all([
+        sys.skills(input.agent),
+        sys.environment(input.model),
+        instruction.system().pipe(Effect.orDie),
+        sys.mcp(input.agent, input.session.permission),
+        MessageV2.toModelMessagesEffect(input.messages, input.model),
+      ])
+      return {
+        tools,
+        modelMessages,
+        system: [...env, ...instructions, ...(mcpInstructions ? [mcpInstructions] : []), ...(skills ? [skills] : [])],
+      }
+    })
+
    const cancel = Effect.fn("SessionPrompt.cancel")(function* (sessionID: SessionID) {
      yield* Effect.logInfo("cancel", { "session.id": sessionID })
      yield* state.cancel(sessionID)
@@ -1153,6 +1192,21 @@ const layer = Layer.effect(
              sessionID,
              auto: task.auto,
              overflow: task.overflow,
+              run: ({ user, agent, model, processor, messages, prompt }) =>
+                Effect.gen(function* () {
+                  const prepared = yield* prepare({ session, agent, model, processor, messages })
+                  return yield* processor.process({
+                    user,
+                    agent,
+                    permission: session.permission,
+                    sessionID,
+                    parentSessionID: session.parentID,
+                    system: prepared.system,
+                    messages: [...prepared.modelMessages, { role: "user", content: prompt }],
+                    tools: prepared.tools,
+                    model,
+                  })
+                }),
            })
            if (result === "stop") break
            continue
@@ -1219,68 +1273,34 @@ const layer = Layer.effect(
            .pipe(Effect.onInterrupt(() => finalizeInterruptedAssistant))

          const outcome: "break" | "continue" = yield* Effect.gen(function* () {
-            const lastUserMsg = msgs.findLast((m) => m.info.role === "user")
-            const bypassAgentCheck = lastUserMsg?.parts.some((p) => p.type === "agent") ?? false
-            const promptOps = yield* ops()
-
-            const tools = yield* SessionTools.resolve({
-              agent,
-              session,
-              model,
-              processor: handle,
-              bypassAgentCheck,
-              messages: msgs,
-              promptOps,
-            }).pipe(
-              Effect.provideService(Plugin.Service, plugin),
-              Effect.provideService(Permission.Service, permission),
-              Effect.provideService(ToolRegistry.Service, registry),
-              Effect.provideService(MCP.Service, mcp),
-              Effect.provideService(Truncate.Service, truncate),
-              Effect.provideService(RuntimeFlags.Service, flags),
-            )
-
-            if (lastUser.format?.type === "json_schema") {
-              tools["StructuredOutput"] = createStructuredOutputTool({
-                schema: lastUser.format.schema,
-                onSuccess(output) {
-                  structured = output
-                },
-              })
-            }
-
            if (step === 1)
              yield* summary.summarize({ sessionID, messageID: lastUser.id }).pipe(Effect.ignore, Effect.forkIn(scope))

            yield* plugin.trigger("experimental.chat.messages.transform", {}, { messages: msgs })

-            const [skills, env, instructions, mcpInstructions, modelMsgs] = yield* Effect.all([
-              sys.skills(agent),
-              sys.environment(model),
-              instruction.system().pipe(Effect.orDie),
-              sys.mcp(agent, session.permission),
-              MessageV2.toModelMessagesEffect(msgs, model),
-            ])
-            const system = [
-              ...env,
-              ...instructions,
-              ...(mcpInstructions ? [mcpInstructions] : []),
-              ...(skills ? [skills] : []),
-            ]
+            const prepared = yield* prepare({ session, agent, model, processor: handle, messages: msgs })
            const format = lastUser.format ?? { type: "text" as const }
-            if (format.type === "json_schema") system.push(STRUCTURED_OUTPUT_SYSTEM_PROMPT)
+            if (format.type === "json_schema") {
+              prepared.system.push(STRUCTURED_OUTPUT_SYSTEM_PROMPT)
+              prepared.tools["StructuredOutput"] = createStructuredOutputTool({
+                schema: format.schema,
+                onSuccess(output) {
+                  structured = output
+                },
+              })
+            }
            const result = yield* handle.process({
              user: lastUser,
              agent,
              permission: session.permission,
              sessionID,
              parentSessionID: session.parentID,
-              system,
+              system: prepared.system,
              messages: [
-                ...modelMsgs,
+                ...prepared.modelMessages,
                ...(isLastStep ? [{ role: "assistant" as const, content: MAX_STEPS_PROMPT }] : []),
              ],
-              tools,
+              tools: prepared.tools,
              model,
              toolChoice: format.type === "json_schema" ? "required" : undefined,
            })
@@ -1,4 +1,4 @@
-You are OpenCode, a coding agent that helps users with software engineering tasks. You are powered by Muse Spark, a large language model trained by Meta MSL.
+You are OpenCode, a coding agent that helps users with software engineering tasks. You are powered by {{MODEL_NAME}}, a large language model trained by Meta MSL.

 Use the instructions below and the tools available to assist the user.

@@ -61,5 +61,5 @@ Use the instructions below and the tools available to assist the user.
 - NEVER use comments as a place for long-winded chain-of-thought. Long thinking texts must be generated as private reasoning. Comments in code must be appropriately concise.

 # User Help & Feedback
- Users can give feedback or report issues at https://github.com/anomalyco/opencode and mention that they are using Meta Muse Spark.
+- Users can give feedback or report issues at https://github.com/anomalyco/opencode and mention that they are using Meta {{MODEL_NAME}}.
 - When users ask directly about OpenCode (eg. "can OpenCode do...", "are you able to do...") or its features (eg. implement a hook, write a slash command, or install an MCP server), use the WebFetch tool to gather information to answer the question from the OpenCode docs at https://opencode.ai/docs.
@@ -25,8 +25,10 @@ export type Retryable = {

 export const RETRY_INITIAL_DELAY = 2000
 export const RETRY_BACKOFF_FACTOR = 2
+export const RETRY_JITTER_FACTOR = 0.25
 export const RETRY_MAX_DELAY_NO_HEADERS = 30_000 // 30 seconds
 export const RETRY_MAX_DELAY = 2_147_483_647 // max 32-bit signed integer for setTimeout
+export const RETRY_MAX_RETRIES = 5

 const RETRYABLE_MESSAGE_PATTERNS = [
  /429|500|502|503|504|524/i,
@@ -41,7 +43,7 @@ function cap(ms: number) {
  return Math.min(ms, RETRY_MAX_DELAY)
 }

-export function delay(attempt: number, error?: SessionV1.APIError) {
+export function delay(attempt: number, error?: SessionV1.APIError, random = Math.random()) {
  if (error) {
    const headers = error.data.responseHeaders
    if (headers) {
@@ -67,11 +69,16 @@ export function delay(attempt: number, error?: SessionV1.APIError) {
        }
      }

-      return cap(RETRY_INITIAL_DELAY * Math.pow(RETRY_BACKOFF_FACTOR, attempt - 1))
+      return cap(exponential(attempt, random))
    }
  }

-  return cap(Math.min(RETRY_INITIAL_DELAY * Math.pow(RETRY_BACKOFF_FACTOR, attempt - 1), RETRY_MAX_DELAY_NO_HEADERS))
+  return cap(Math.min(exponential(attempt, random), RETRY_MAX_DELAY_NO_HEADERS))
+}
+
+function exponential(attempt: number, random: number) {
+  const base = RETRY_INITIAL_DELAY * Math.pow(RETRY_BACKOFF_FACTOR, attempt - 1)
+  return Math.ceil(base + base * RETRY_JITTER_FACTOR * random)
 }

 export function retryable(error: Err, provider: string) {
@@ -182,6 +189,7 @@ export function policy(opts: {
      const error = opts.parse(meta.input)
      const retry = retryable(error, opts.provider)
      if (!retry) return Cause.done(meta.attempt)
+      if (meta.attempt > RETRY_MAX_RETRIES) return Cause.done(meta.attempt)
      return Effect.gen(function* () {
        const wait = delay(meta.attempt, SessionV1.APIError.isInstance(error) ? error : undefined)
        const now = yield* Clock.currentTimeMillis
@@ -25,7 +25,10 @@ import { MCP } from "@/mcp"
 import { PermissionV1 } from "@opencode-ai/core/v1/permission"

 export function provider(model: Provider.Model) {
-  if (model.api.id.includes("muse-spark")) return [PROMPT_META]
+  if (model.api.id.includes("muse")) {
+    const name = model.api.id.includes("muse-glimmer") ? "Muse Glimmer" : "Muse Spark"
+    return [PROMPT_META.replaceAll("{{MODEL_NAME}}", name)]
+  }
  if (model.api.id.includes("gpt-4") || model.api.id.includes("o1") || model.api.id.includes("o3"))
    return [PROMPT_BEAST]
  if (model.api.id.includes("gpt")) {
@@ -37,7 +40,11 @@ export function provider(model: Provider.Model) {
  if (model.api.id.includes("gemini-")) return [PROMPT_GEMINI]
  if (model.api.id.includes("claude")) return [PROMPT_ANTHROPIC]
  if (model.api.id.toLowerCase().includes("trinity")) return [PROMPT_TRINITY]
-  if (model.api.id.toLowerCase().includes("kimi")) return [PROMPT_KIMI]
+  if (
+    model.api.id.toLowerCase().includes("kimi") ||
+    ["kimi-for-coding", "moonshotai", "moonshotai-cn"].includes(model.providerID)
+  )
+    return [PROMPT_KIMI]
  return [PROMPT_DEFAULT]
 }

@@ -52,7 +52,7 @@ it.instance("returns default native agents when no config", () =>
    expect(names).toContain("plan")
    expect(names).toContain("general")
    expect(names).toContain("explore")
-    expect(names).toContain("compaction")
+    expect(names).not.toContain("compaction")
    expect(names).toContain("title")
    expect(names).toContain("summary")
  }),
@@ -170,17 +170,34 @@ it.instance("general agent denies todo tools", () =>
  }),
 )

-it.instance("compaction agent denies all permissions", () =>
+it.instance("compaction agent is absent without configuration", () =>
  Effect.gen(function* () {
    const compaction = yield* load((svc) => svc.get("compaction"))
-    expect(compaction).toBeDefined()
-    expect(compaction?.hidden).toBe(true)
-    expect(evalPerm(compaction, "bash")).toBe("deny")
-    expect(evalPerm(compaction, "edit")).toBe("deny")
-    expect(evalPerm(compaction, "read")).toBe("deny")
+    expect(compaction).toBeUndefined()
  }),
 )

+it.instance(
+  "configured compaction agent keeps native defaults",
+  () =>
+    Effect.gen(function* () {
+      const compaction = yield* load((svc) => svc.get("compaction"))
+      expect(compaction).toBeDefined()
+      expect(compaction?.native).toBe(true)
+      expect(compaction?.hidden).toBe(true)
+      expect(evalPerm(compaction, "bash")).toBe("deny")
+      expect(evalPerm(compaction, "edit")).toBe("deny")
+      expect(evalPerm(compaction, "read")).toBe("deny")
+    }),
+  {
+    config: {
+      agent: {
+        compaction: {},
+      },
+    },
+  },
+)
+
 it.instance(
  "custom agent from config creates new agent",
  () =>
@@ -710,6 +727,9 @@ it.instance(
  {
    config: {
      default_agent: "compaction",
+      agent: {
+        compaction: {},
+      },
    },
  },
 )
@@ -597,12 +597,12 @@ accountTokenIt.instance("resolves env templates in account config with account t
  }),
 )

-it.instance("validates config schema and throws on invalid fields", () =>
+it.instance("validates config schema and throws on invalid values", () =>
  Effect.gen(function* () {
    const test = yield* TestInstance
    yield* writeConfigEffect(test.directory, {
      $schema: "https://opencode.ai/config.json",
-      invalid_field: "should cause error",
+      model: 42,
    })
    const exit = yield* Config.use.get().pipe(Effect.exit)
    expect(Exit.isFailure(exit)).toBe(true)
@@ -1331,7 +1331,7 @@ it.instance("permission config preserves user key order", () =>
  }),
 )

-test("config parser preserves permission order while rejecting unknown top-level keys", () => {
+test("config parser preserves permission order while ignoring unknown top-level keys", () => {
  const config = ConfigParse.schema(
    ConfigV1.Info,
    {
@@ -1340,18 +1340,13 @@ test("config parser preserves permission order while rejecting unknown top-level
        "*": "deny",
        edit: "ask",
      },
+      plugins: ["example"],
    },
    "test",
  )

  expect(Object.keys(config.permission!)).toEqual(["bash", "*", "edit"])
-  try {
-    ConfigParse.schema(ConfigV1.Info, { invalid_field: true }, "test")
-    throw new Error("expected config parse to fail")
-  } catch (err) {
-    const error = err as { data?: { issues?: Array<{ code?: string; keys?: string[]; path?: string[] }> } }
-    expect(error.data?.issues?.[0]).toMatchObject({ code: "unrecognized_keys", keys: ["invalid_field"], path: [] })
-  }
+  expect(config).not.toHaveProperty("plugins")
 })

 // MCP config merging tests
@@ -187,6 +187,74 @@ test("converts Copilot AIC token prices to USD per million tokens", async () =>
  expect(models["ignored-non-chat-record"]).toBeUndefined()
 })

+test("detects PDF input support when vision and media type are advertised", async () => {
+  globalThis.fetch = mock(() =>
+    Promise.resolve(
+      new Response(
+        JSON.stringify({
+          data: [
+            {
+              model_picker_enabled: true,
+              id: "pdf-model",
+              name: "PDF Model",
+              version: "pdf-model-2026-06-01",
+              capabilities: {
+                family: "pdf-model",
+                limits: {
+                  max_context_window_tokens: 128000,
+                  max_output_tokens: 16384,
+                  max_prompt_tokens: 128000,
+                  vision: {
+                    max_prompt_image_size: 10000000,
+                    max_prompt_images: 10,
+                    supported_media_types: ["application/pdf"],
+                  },
+                },
+                supports: {
+                  streaming: true,
+                  vision: true,
+                  tool_calls: true,
+                },
+              },
+            },
+            {
+              model_picker_enabled: true,
+              id: "vision-only-model",
+              name: "Vision Only Model",
+              version: "vision-only-model-2026-06-01",
+              capabilities: {
+                family: "vision-only-model",
+                limits: {
+                  max_context_window_tokens: 128000,
+                  max_output_tokens: 16384,
+                  max_prompt_tokens: 128000,
+                  vision: {
+                    max_prompt_image_size: 10000000,
+                    max_prompt_images: 10,
+                    supported_media_types: ["image/png"],
+                  },
+                },
+                supports: {
+                  streaming: true,
+                  vision: true,
+                  tool_calls: true,
+                },
+              },
+            },
+          ],
+        }),
+        { status: 200 },
+      ),
+    ),
+  ) as unknown as typeof fetch
+
+  const models = (await CopilotModels.get("https://api.githubcopilot.com")).models
+  const model = models["pdf-model"]
+
+  expect(model.capabilities.input.pdf).toBe(true)
+  expect(models["vision-only-model"].capabilities.input.pdf).toBe(false)
+})
+
 test("uses zero cost when Copilot reports a zero billing batch size", async () => {
  globalThis.fetch = mock(() =>
    Promise.resolve(
@@ -1548,6 +1548,33 @@ test("models.dev reasoning options replace generated variants and unsupported to
  expect(models["gemini-3-pro-fast"].variants).toEqual(models.override.variants)
 })

+test("MERGE Gateway exposes declared effort variants without model-specific handling", () => {
+  const provider = {
+    id: "merge-gateway",
+    name: "MERGE Gateway",
+    env: ["MERGE_GATEWAY_API_KEY"],
+    npm: "merge-gateway-ai-sdk-provider",
+    models: {
+      "openai/gpt-5.6-sol": {
+        id: "openai/gpt-5.6-sol",
+        name: "GPT-5.6 Sol",
+        reasoning: true,
+        reasoning_options: [{ type: "effort", values: ["none", "low", "medium", "high", "xhigh", "max"] }],
+        limit: { context: 128_000, output: 64_000 },
+      },
+    },
+  } as unknown as ModelsDev.Provider
+
+  expect(Provider.fromModelsDevProvider(provider).models["openai/gpt-5.6-sol"].variants).toEqual({
+    none: { reasoningEffort: "none" },
+    low: { reasoningEffort: "low" },
+    medium: { reasoningEffort: "medium" },
+    high: { reasoningEffort: "high" },
+    xhigh: { reasoningEffort: "xhigh" },
+    max: { reasoningEffort: "max" },
+  })
+})
+
 test("public provider info omits invalid models", () => {
  const provider = Provider.fromModelsDevProvider({
    id: "test",
@@ -3272,6 +3272,37 @@ describe("ProviderTransform sampling defaults - Gemini", () => {
  })
 })

+describe("ProviderTransform sampling defaults - DeepSeek", () => {
+  const model = (providerID: string, id: string) =>
+    ({
+      id: `${providerID}/${id}`,
+      providerID,
+      api: { id },
+    }) as any
+
+  test.each([
+    ["deepseek", "deepseek-v4-flash"],
+    ["opencode", "deepseek-v4-flash"],
+    ["opencode-go", "deepseek-v4-flash"],
+    ["openrouter", "deepseek/deepseek-v4-flash-0731"],
+    ["ollama-cloud", "deepseek-v4-flash:0731"],
+  ])("defaults top_p for %s/%s", (providerID, id) => {
+    expect(ProviderTransform.temperature(model(providerID, id))).toBeUndefined()
+    expect(ProviderTransform.topP(model(providerID, id))).toBe(0.95)
+    expect(ProviderTransform.topK(model(providerID, id))).toBeUndefined()
+  })
+
+  test.each([
+    ["openrouter", "deepseek/deepseek-v4-flash"],
+    ["vercel", "deepseek/deepseek-v4-flash"],
+    ["custom", "deepseek-ai/DeepSeek-V4-Flash"],
+  ])("preserves legacy defaults for %s/%s", (providerID, id) => {
+    expect(ProviderTransform.temperature(model(providerID, id))).toBeUndefined()
+    expect(ProviderTransform.topP(model(providerID, id))).toBeUndefined()
+    expect(ProviderTransform.topK(model(providerID, id))).toBeUndefined()
+  })
+})
+
 describe("ProviderTransform.reasoningVariants", () => {
  const model = (reasoning_options: ModelsDev.Model["reasoning_options"]) => ({ reasoning_options }) as ModelsDev.Model
  const target = (npm: string, id = "test-model") =>
@@ -3339,6 +3370,7 @@ describe("ProviderTransform.reasoningVariants", () => {
    ["@ai-sdk/togetherai", { reasoningEffort: "high" }],
    ["venice-ai-sdk-provider", { reasoningEffort: "high" }],
    ["ai-gateway-provider", { reasoningEffort: "high" }],
+    ["merge-gateway-ai-sdk-provider", { reasoningEffort: "high" }],
    ["@ai-sdk/amazon-bedrock", { reasoningConfig: { type: "enabled", maxReasoningEffort: "high" } }],
  ])("converts effort for %s", (npm, expected, ...args) => {
    const id = args[0] as string | undefined
@@ -5524,6 +5556,25 @@ describe("ProviderTransform.providerOptions - ai-gateway-provider", () => {
  })
 })

+describe("ProviderTransform.providerOptions - merge-gateway-ai-sdk-provider", () => {
+  const model = {
+    id: "merge-gateway/openai/gpt-5.6-sol",
+    providerID: "merge-gateway",
+    api: {
+      id: "openai/gpt-5.6-sol",
+      url: "https://api-gateway.merge.dev/v1/ai-sdk",
+      npm: "merge-gateway-ai-sdk-provider",
+    },
+    capabilities: { reasoning: true },
+  } as any
+
+  test("routes normalized effort under the adapter's mergeGateway key", () => {
+    expect(ProviderTransform.providerOptions(model, { reasoningEffort: "high" })).toEqual({
+      mergeGateway: { reasoningEffort: "high" },
+    })
+  })
+})
+
 describe("ProviderTransform.options - kimi family adaptive thinking", () => {
  const createModel = (overrides: Record<string, any> = {}) =>
    ({
@@ -161,6 +161,21 @@ describe("HttpApi instance context middleware", () => {
    }),
  )

+  it.live("persists the routed project while loading instance context", () =>
+    Effect.gen(function* () {
+      const dir = yield* tmpdirScoped({ git: true })
+      const project = yield* Project.Service
+      yield* serveProbe()
+
+      const response = yield* HttpClient.get(`/probe?directory=${encodeURIComponent(dir)}`)
+
+      expect(response.status).toBe(200)
+      const saved = (yield* project.list()).find((item) => item.worktree === dir)
+      expect(saved).toBeDefined()
+      expect(saved?.id).not.toBe("global")
+    }),
+  )
+
  it.live("falls back to the raw directory when URI decoding fails", () =>
    Effect.gen(function* () {
      yield* serveProbe()
@@ -222,6 +222,11 @@ function cfg(compaction?: ConfigV1.Info["compaction"]) {
  return Layer.succeed(Config.Service, TestConfig.make({ get: () => Effect.succeed({ ...base, compaction }) }))
 }

+function cfgAgent(agent: NonNullable<ConfigV1.Info["agent"]>) {
+  const config = Schema.decodeUnknownSync(ConfigV1.Info)({ agent }) as ConfigV1.Info
+  return Layer.succeed(Config.Service, TestConfig.make({ get: () => Effect.succeed(config) }))
+}
+
 const defaultProvider = wide()
 const compactionTestNode = LayerNode.group([
  SessionCompaction.node,
@@ -365,6 +370,20 @@ function autocontinue(enabled: boolean) {
  })
 }

+function compactionContext(context: string) {
+  return Layer.mock(Plugin.Service)({
+    trigger: <Name extends string, Input, Output>(name: Name, _input: Input, output: Output) => {
+      if (name !== "experimental.session.compacting") return Effect.succeed(output)
+      return Effect.sync(() => {
+        ;(output as { context: string[] }).context.push(context)
+        return output
+      })
+    },
+    list: () => Effect.succeed([]),
+    init: () => Effect.void,
+  })
+}
+
 describe("session.compaction.isOverflow", () => {
  it.live(
    "returns true when token count exceeds usable context",
@@ -798,6 +817,75 @@ describe("session.compaction.prune", () => {
 })

 describe("session.compaction.process", () => {
+  itCompaction.instance(
+    "inherits the last agent when compaction is not configured",
+    () => {
+      const stub = llm()
+      let captured: LLM.StreamInput | undefined
+      stub.push(reply("summary", (input) => (captured = input)))
+      return Effect.gen(function* () {
+        const ssn = yield* SessionNs.Service
+        const session = yield* ssn.create({})
+        const msg = yield* createUserMessage(session.id, "hello")
+        const msgs = yield* ssn.messages({ sessionID: session.id })
+
+        yield* SessionCompaction.use.process({
+          parentID: msg.id,
+          messages: msgs,
+          sessionID: session.id,
+          auto: false,
+        })
+
+        expect(captured?.agent.name).toBe("build")
+        expect(captured?.agent.prompt).toBe("custom build prompt")
+        expect(captured?.agent.temperature).toBe(0.37)
+        expect(String(captured?.model.providerID)).toBe("test")
+        expect(String(captured?.model.id)).toBe("test-model")
+      }).pipe(
+        withCompaction({
+          llm: stub.llmLayer,
+          config: cfgAgent({
+            build: { prompt: "custom build prompt", temperature: 0.37, model: "missing/missing" },
+          }),
+        }),
+      )
+    },
+    { git: true },
+  )
+
+  itCompaction.instance(
+    "preserves an explicitly configured compaction agent",
+    () => {
+      const stub = llm()
+      let captured: LLM.StreamInput | undefined
+      stub.push(reply("summary", (input) => (captured = input)))
+      return Effect.gen(function* () {
+        const ssn = yield* SessionNs.Service
+        const session = yield* ssn.create({})
+        const msg = yield* createUserMessage(session.id, "hello")
+        const msgs = yield* ssn.messages({ sessionID: session.id })
+
+        yield* SessionCompaction.use.process({
+          parentID: msg.id,
+          messages: msgs,
+          sessionID: session.id,
+          auto: false,
+        })
+
+        expect(captured?.agent.name).toBe("compaction")
+        expect(captured?.agent.prompt).toBe("custom compaction prompt")
+        expect(captured?.agent.temperature).toBe(0.81)
+        expect(captured?.agent.hidden).toBe(true)
+      }).pipe(
+        withCompaction({
+          llm: stub.llmLayer,
+          config: cfgAgent({ compaction: { prompt: "custom compaction prompt", temperature: 0.81 } }),
+        }),
+      )
+    },
+    { git: true },
+  )
+
  it.instance(
    "throws when parent is not a user message",
    Effect.gen(function* () {
@@ -1389,11 +1477,21 @@ describe("session.compaction.process", () => {
        const captured = JSON.stringify(messages)
        expect(messages).toHaveLength(1)
        expect(messages[0]?.role).toBe("user")
+        expect(captured).toContain("Here is the conversation so far:")
+        expect(captured).toContain("<conversation>")
+        expect(captured.indexOf("[User]: older context")).toBeLessThan(
+          captured.indexOf("Create a new anchored summary"),
+        )
        expect(captured).toContain("[User]: older context")
        expect(captured).not.toContain("keep this turn")
        expect(captured).not.toContain("and this one too")
        expect(captured).not.toContain("What did we do so far?")
-      }).pipe(withCompaction({ llm: stub.llmLayer }))
+      }).pipe(
+        withCompaction({
+          llm: stub.llmLayer,
+          config: cfg({ tail_turns: 2, preserve_recent_tokens: 10_000 }),
+        }),
+      )
    },
    { git: true },
  )
@@ -1430,9 +1528,11 @@ describe("session.compaction.process", () => {
        expect(parent).toBeTruthy()
        yield* SessionCompaction.use.process({ parentID: parent!, messages: msgs, sessionID: session.id, auto: false })

-        expect(captured).toContain("<previous-summary>")
+        expect(captured).toContain("<prior-summary>")
        expect(captured).toContain("summary one")
        expect(captured.match(/summary one/g)?.length).toBe(1)
+        expect(captured.indexOf("latest turn")).toBeLessThan(captured.indexOf("<prior-summary>"))
+        expect(captured).toContain("summary of the conversation before the <conversation> above")
        expect(captured).toContain("## Important Details")
        expect(captured).toContain("## Work State")
      }).pipe(withCompaction({ llm: stub.llmLayer }))
@@ -1440,6 +1540,49 @@ describe("session.compaction.process", () => {
    { git: true },
  )

+  itCompaction.instance(
+    "keeps plugin context outside the serialized conversation",
+    () => {
+      const stub = llm()
+      let captured = ""
+      stub.push(
+        reply("summary", (input) => {
+          captured = JSON.stringify(input.messages)
+        }),
+      )
+
+      return Effect.gen(function* () {
+        const ssn = yield* SessionNs.Service
+        const session = yield* ssn.create({})
+        yield* createUserMessage(session.id, "older context")
+        yield* createUserMessage(session.id, "keep this turn")
+        yield* createUserMessage(session.id, "and this one too")
+        yield* createCompactionMarker(session.id)
+
+        const msgs = yield* ssn.messages({ sessionID: session.id })
+        const parent = msgs.at(-1)?.info.id
+        expect(parent).toBeTruthy()
+        yield* SessionCompaction.use.process({
+          parentID: parent!,
+          messages: msgs,
+          sessionID: session.id,
+          auto: false,
+        })
+
+        expect(captured).toContain("Prioritize unresolved migration details")
+        expect(captured.indexOf("</conversation>")).toBeLessThan(
+          captured.indexOf("Prioritize unresolved migration details"),
+        )
+      }).pipe(
+        withCompaction({
+          llm: stub.llmLayer,
+          plugin: compactionContext("Prioritize unresolved migration details"),
+        }),
+      )
+    },
+    { git: true },
+  )
+
  itCompaction.instance(
    "serializes repeated compaction history as one user message",
    () => {
@@ -444,6 +444,34 @@ const boot = Effect.fn("test.boot")(function* (input?: { title?: string }) {

 // Loop semantics

+it.instance("default compaction reuses the normal system and tools", () =>
+  Effect.gen(function* () {
+    const { llm } = yield* useServerConfig(providerCfg)
+    const { prompt, chat } = yield* boot()
+    const compact = yield* SessionCompaction.Service
+    yield* user(chat.id, "cache prefix message")
+    yield* llm.text("normal response")
+    yield* prompt.loop({ sessionID: chat.id })
+
+    yield* compact.create({ sessionID: chat.id, agent: "build", model: ref, auto: false })
+    yield* llm.text("compacted summary")
+    yield* prompt.loop({ sessionID: chat.id })
+
+    const inputs = yield* llm.inputs
+    expect(inputs).toHaveLength(2)
+    expect(inputs[1]?.tools).toEqual(inputs[0]?.tools)
+    const normalMessages = Array.isArray(inputs[0]?.messages) ? inputs[0].messages : []
+    const compactMessages = Array.isArray(inputs[1]?.messages) ? inputs[1].messages : []
+    const system = (message: unknown) =>
+      typeof message === "object" && message !== null && "role" in message && message.role === "system"
+    expect(compactMessages.find(system)).toEqual(normalMessages.find(system))
+    expect(JSON.stringify(compactMessages)).toContain("cache prefix message")
+    expect(JSON.stringify(compactMessages)).toContain("normal response")
+    expect(JSON.stringify(compactMessages)).toContain("Create a new anchored summary")
+    expect(JSON.stringify(compactMessages)).not.toContain("[User]: cache prefix message")
+  }),
+)
+
 noLLMServer.instance(
  "loop exits immediately when last assistant has stop finish",
  () =>
@@ -35,10 +35,18 @@ function wrap(message: unknown): ReturnType<NamedError["toObject"]> {
 describe("session.retry.delay", () => {
  test("caps delay at 30 seconds when headers missing", () => {
    const error = apiError()
-    const delays = Array.from({ length: 10 }, (_, index) => SessionRetry.delay(index + 1, error))
+    const delays = Array.from({ length: 10 }, (_, index) => SessionRetry.delay(index + 1, error, 0))
    expect(delays).toStrictEqual([2000, 4000, 8000, 16000, 30000, 30000, 30000, 30000, 30000, 30000])
  })

+  test("adds jitter to exponential delays", () => {
+    const error = apiError()
+    expect(SessionRetry.delay(1, error, 0)).toBe(2000)
+    expect(SessionRetry.delay(1, error, 1)).toBe(2500)
+    expect(SessionRetry.delay(4, error, 1)).toBe(20000)
+    expect(SessionRetry.delay(5, error, 1)).toBe(30000)
+  })
+
  test("prefers retry-after-ms when shorter than exponential", () => {
    const error = apiError({ "retry-after-ms": "1500" })
    expect(SessionRetry.delay(4, error)).toBe(1500)
@@ -59,18 +67,18 @@ describe("session.retry.delay", () => {

  test("ignores invalid retry hints", () => {
    const error = apiError({ "retry-after": "not-a-number" })
-    expect(SessionRetry.delay(1, error)).toBe(2000)
+    expect(SessionRetry.delay(1, error, 0)).toBe(2000)
  })

  test("ignores malformed date retry hints", () => {
    const error = apiError({ "retry-after": "Invalid Date String" })
-    expect(SessionRetry.delay(1, error)).toBe(2000)
+    expect(SessionRetry.delay(1, error, 0)).toBe(2000)
  })

  test("ignores past date retry hints", () => {
    const pastDate = new Date(Date.now() - 5000).toUTCString()
    const error = apiError({ "retry-after": pastDate })
-    expect(SessionRetry.delay(1, error)).toBe(2000)
+    expect(SessionRetry.delay(1, error, 0)).toBe(2000)
  })

  test("uses retry-after values even when exceeding 10 minutes with headers", () => {
@@ -115,6 +123,29 @@ describe("session.retry.delay", () => {
      })
    }),
  )
+
+  it.instance("policy stops after five retries", () =>
+    Effect.gen(function* () {
+      const attempts: number[] = []
+      const error = apiError({ "retry-after-ms": "0" })
+      const step = yield* Schedule.toStepWithMetadata(
+        SessionRetry.policy({
+          provider: "test",
+          parse: Schema.decodeUnknownSync(SessionV1.APIError.Schema),
+          set: (info) =>
+            Effect.sync(() => {
+              attempts.push(info.attempt)
+            }),
+        }),
+      )
+
+      yield* Effect.forEach(Array.from({ length: SessionRetry.RETRY_MAX_RETRIES + 1 }), () =>
+        Effect.ignore(step(error)),
+      )
+
+      expect(attempts).toStrictEqual([1, 2, 3, 4, 5])
+    }),
+  )
 })

 describe("session.retry.retryable", () => {
@@ -85,9 +85,28 @@ const it = testEffect(

 describe("session.system", () => {
  test("selects the Meta prompt for Muse Spark model IDs", () => {
-    expect(SystemPrompt.provider({ api: { id: "meta/muse-spark-preview" } } as Provider.Model)[0]).toContain(
-      "Meta Muse Spark",
-    )
+    for (const id of ["meta/muse-spark-preview", "muse-spark-1.1", "muse-spark-1.2"]) {
+      const prompt = SystemPrompt.provider({ api: { id } } as Provider.Model)[0]
+      expect(prompt).toContain("powered by Muse Spark,")
+      expect(prompt).toContain("using Meta Muse Spark.")
+      expect(prompt).not.toContain("{{MODEL_NAME}}")
+    }
+  })
+
+  test("selects the Meta prompt for Muse Glimmer model IDs", () => {
+    for (const id of ["meta/muse-glimmer", "meta/muse-glimmer-30b", "muse-glimmer-30b"]) {
+      const prompt = SystemPrompt.provider({ api: { id } } as Provider.Model)[0]
+      expect(prompt).toContain("powered by Muse Glimmer,")
+      expect(prompt).toContain("using Meta Muse Glimmer.")
+      expect(prompt).not.toContain("{{MODEL_NAME}}")
+    }
+  })
+
+  test("selects the Kimi prompt for official provider model IDs", () => {
+    for (const providerID of ["kimi-for-coding", "moonshotai", "moonshotai-cn"]) {
+      const prompt = SystemPrompt.provider({ providerID, api: { id: "k3" } } as Provider.Model)[0]
+      expect(prompt).toContain("# Prompt and Tool Use")
+    }
  })

  it.effect("skills output is sorted by name and stable across calls", () =>
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/plugin",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/sdk",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/server",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/session-ui",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -5,6 +5,10 @@ import { PreloadMultiFileDiffResult } from "@pierre/diffs/ssr"

 export type NormalizedProviderListResponse = {
  all: Map<string, Provider>
+  defaultModel?: {
+    providerID: string
+    modelID: string
+  } | null
  default: {
    [key: string]: string
  }
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/slack",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "scripts": {
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/stats-app",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/stats-core",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -12,6 +12,7 @@
    "./database": "./src/database.ts",
    "./database/*": "./src/database/*.ts",
    "./domain/*": "./src/domain/*.ts",
+    "./r2-sql": "./src/r2-sql.ts",
    "./runtime": "./src/runtime.ts",
    "./stat-sync": "./src/stat-sync.ts"
  },
@@ -9,7 +9,7 @@ import { Effect, Layer } from "effect"
 import * as Context from "effect/Context"
 import { Resource } from "sst/resource"

-const ATHENA_MAX_POLL_ATTEMPTS = 300
+const ATHENA_MAX_POLL_ATTEMPTS = 900
 const ATHENA_PAGE_SIZE = 1000

 export type AthenaData = Record<string, string>
@@ -1,5 +1,5 @@
 import { describe, expect, test } from "bun:test"
-import { toGeoAggregate, toModelAggregate, toProviderAggregate } from "./inference"
+import { buildStatsQueries, toGeoAggregate, toModelAggregate, toProviderAggregate } from "./inference"
 import { modelAuthor, normalizeInferenceModel, statModel, statProvider } from "./model-normalization"

 describe("inference stat normalization", () => {
@@ -82,6 +82,55 @@ describe("inference stat normalization", () => {
      }),
    ).toMatchObject([{ period_key: "2026-W20" }])
  })
+
+  test("builds bounded R2 SQL queries for each day and week", () => {
+    const queries = buildStatsQueries(new Date("2026-08-10T00:00:00.000Z"), new Date("2026-08-12T12:00:00.000Z"), {
+      namespace: "inference",
+      table: "generation",
+      dataset: "zen",
+    })
+
+    expect(queries).toHaveLength(8)
+    expect(queries[0]).toContain("'week' AS grain")
+    expect(queries[0]).toContain("'2026-W33' AS period_key")
+    expect(queries[2]).toContain("'2026-08-10' AS period_key")
+    expect(queries[6]).toContain("'2026-08-12' AS period_key")
+    expect(queries[0]).toContain('FROM "inference"."generation"')
+    expect(queries[0]).toContain("event_type = 'generation.completed'")
+    expect(queries[0]).toContain("product = 'go'")
+    expect(queries[0]).toContain("LIMIT 10000")
+    expect(queries[0]).toContain("approx_distinct(session) AS sessions")
+    expect(queries[1]).toContain("'geo_model' ELSE 'geo'")
+    expect(queries[1]).toContain("0 AS sessions")
+  })
+
+  test("aligns periods to UTC calendar boundaries", () => {
+    const queries = buildStatsQueries(new Date("2026-06-17T15:56:00.000Z"), new Date("2026-06-19T15:56:00.000Z"), {
+      namespace: "inference",
+      table: "generation",
+      dataset: "zen",
+    })
+
+    expect(queries).toHaveLength(8)
+    expect(queries[0]).toContain("'2026-W25' AS period_key")
+    expect(queries[0]).toContain("started_at >= '2026-06-15T00:00:00.000Z'")
+    expect(queries[2]).toContain("'2026-06-17' AS period_key")
+    expect(queries[2]).toContain("started_at >= '2026-06-17T00:00:00.000Z'")
+    expect(queries[2]).toContain("started_at < '2026-06-18T00:00:00.000Z'")
+    expect(queries[6]).toContain("'2026-06-19' AS period_key")
+    expect(queries[6]).toContain("started_at < '2026-06-19T15:56:00.000Z'")
+  })
+
+  test("uses an exclusive live and legacy source handoff", () => {
+    const [query] = buildStatsQueries(new Date("2026-08-11T00:00:00.000Z"), new Date("2026-08-12T00:00:00.000Z"), {
+      namespace: "inference",
+      table: "generation",
+      dataset: "zen",
+    })
+
+    expect(query).toContain("(source = 'inference-legacy' AND started_at < '2026-08-11T10:57:48.186Z')")
+    expect(query).toContain("(source = 'inference' AND started_at >= '2026-08-11T10:57:48.186Z')")
+  })
 })

 function aggregate(model: string, provider: string) {
@@ -1,5 +1,5 @@
 import { Resource } from "sst/resource"
-import type { AthenaData } from "../athena"
+import type { R2SqlData } from "../r2-sql"
 import type { GeoStatAggregate } from "./geo"
 import type { ModelStatAggregate } from "./model"
 import {
@@ -10,25 +10,80 @@ import {
  statProvider,
 } from "./model-normalization"
 import type { ProviderStatAggregate } from "./provider"
-import { normalizeCountry, normalizeTier, type StatBaseAggregate } from "./stat"
+import {
+  normalizeCountry,
+  normalizeTier,
+  periodKeyFor,
+  startOfIsoWeek,
+  startOfUtcDay,
+  type StatBaseAggregate,
+} from "./stat"

 export type StatDimension = "model" | "provider" | "geo" | "geo_model"
+export type StatsQuerySource = { namespace: string; table: string; dataset: string }
+type StatsQueryFamily = "usage" | "geo"

-// All stat dimensions and both grains are computed in one query via GROUPING SETS so
-// the source table is scanned once per sync pass; separate queries per dimension (and
-// the previous weekly/daily UNION ALL) each re-scanned the same events.
-export function buildStatsQuery(periodStart: Date, periodEnd: Date) {
-  const periodStartValue = sqlString(periodStart.toISOString())
-  const periodEndValue = sqlString(periodEnd.toISOString())
-  const periodStartDateValue = sqlString(periodStart.toISOString().slice(0, 10))
-  const periodEndDateValue = sqlString(periodEnd.toISOString().slice(0, 10))
-  const sourceTable = [Resource.InferenceEvent.catalog, Resource.InferenceEvent.database, Resource.InferenceEvent.table]
-    .map(sqlIdentifier)
-    .join(".")
+const DAY_MS = 86_400_000
+const WEEK_MS = 7 * DAY_MS
+// The typed production stream began before the legacy backfill's original end
+// boundary. Use one exclusive handoff so the overlapping rows are never counted
+// from both sources.
+const LIVE_SOURCE_START = "2026-08-11T10:57:48.186Z"
+
+// R2 SQL limits result sets to 10,000 rows and does not support OFFSET. Two
+// queries per day/week keep each result bounded and avoid combining the costly
+// distinct user/session aggregates with the high-cardinality geo dimensions.
+export function buildStatsQueries(periodStart: Date, periodEnd: Date, input?: StatsQuerySource) {
+  const source = input ?? {
+    namespace: Resource.R2Sql.namespace,
+    table: Resource.R2Sql.table,
+    dataset: Resource.StatsSyncConfig.dataset,
+  }
+  return [...statPeriods("week", periodStart, periodEnd), ...statPeriods("day", periodStart, periodEnd)].flatMap(
+    (period) => [buildStatsQuery(period, source, "usage"), buildStatsQuery(period, source, "geo")],
+  )
+}
+
+function buildStatsQuery(
+  period: { grain: "day" | "week"; key: string; start: Date; end: Date },
+  source: StatsQuerySource,
+  family: StatsQueryFamily,
+) {
+  const periodStartValue = sqlString(period.start.toISOString())
+  const periodEndValue = sqlString(period.end.toISOString())
+  const ingestEndValue = sqlString(new Date(period.end.getTime() + DAY_MS).toISOString())
+  const sourceTable = [source.namespace, source.table].map(sqlIdentifier).join(".")
+  const dimensions =
+    family === "usage"
+      ? `CASE WHEN grouping(model) = 0 THEN 'model' ELSE 'provider' END AS dimension,
+  tier,
+  provider,
+  CASE WHEN grouping(model) = 0 THEN model END AS model,
+  CASE WHEN grouping(model) = 0 THEN COALESCE(MAX(NULLIF(provider_model, '')), '') END AS provider_model,
+  null AS country,
+  null AS continent`
+      : `CASE WHEN grouping(model) = 0 THEN 'geo_model' ELSE 'geo' END AS dimension,
+  tier,
+  CASE WHEN grouping(model) = 0 THEN provider ELSE 'all' END AS provider,
+  CASE WHEN grouping(model) = 0 THEN model ELSE 'all' END AS model,
+  null AS provider_model,
+  country,
+  COALESCE(MAX(NULLIF(continent, '')), '') AS continent`
+  const distinctColumns =
+    family === "usage"
+      ? `approx_distinct(session) AS sessions,
+    approx_distinct(user_key) AS unique_users`
+      : `0 AS sessions,
+    0 AS unique_users`
+  const groupingSets =
+    family === "usage"
+      ? `(tier, provider, model),
+  (tier, provider)`
+      : `(tier, country),
+  (tier, provider, model, country)`
  const aggregateColumns = `
-    COUNT(DISTINCT session) AS sessions,
+    ${distinctColumns},
    COUNT(*) AS requests,
-    COUNT(DISTINCT user_key) AS unique_users,
    COALESCE(SUM(tokens_input), 0) AS input_tokens,
    COALESCE(SUM(tokens_output), 0) AS output_tokens,
    COALESCE(SUM(tokens_reasoning), 0) AS reasoning_tokens,
@@ -38,65 +93,61 @@ export function buildStatsQuery(periodStart: Date, periodEnd: Date) {
    COALESCE(SUM(cost_output_microcents), 0) AS output_cost_microcents,
    COALESCE(SUM(cost_total_microcents), 0) AS total_cost_microcents,
    AVG(duration_ms) AS avg_duration_ms,
-    approx_percentile(CAST(duration_ms AS double), 0.5) AS p50_duration_ms,
-    approx_percentile(CAST(duration_ms AS double), 0.95) AS p95_duration_ms,
+    null AS p50_duration_ms,
+    null AS p95_duration_ms,
    AVG(ttfb_ms) AS avg_ttfb_ms,
-    approx_percentile(CAST(ttfb_ms AS double), 0.5) AS p50_ttfb_ms,
-    approx_percentile(CAST(ttfb_ms AS double), 0.95) AS p95_ttfb_ms,
+    null AS p50_ttfb_ms,
+    null AS p95_ttfb_ms,
    AVG(output_tps) AS avg_output_tps,
-    SUM(CASE WHEN status >= 200 AND status < 400 THEN 1 ELSE 0 END) AS success_count,
-    SUM(CASE WHEN status >= 400 THEN 1 ELSE 0 END) AS error_count,
+    SUM(CASE WHEN outcome = 'succeeded' THEN 1 ELSE 0 END) AS success_count,
+    SUM(CASE WHEN outcome = 'failed' THEN 1 ELSE 0 END) AS error_count,
    COUNT(*) AS sample_count`

  return `
 WITH normalized AS (
  SELECT
-    from_iso8601_timestamp(event_timestamp) AS event_time,
-    model AS raw_model,
-    ${statModelSql("model", "provider_model")} AS model,
-    COALESCE(NULLIF(provider_model, ''), '') AS provider_model,
-    COALESCE(NULLIF(provider, ''), '') AS raw_provider,
-    UPPER(COALESCE(NULLIF(cf_country, ''), 'ZZ')) AS country,
-    COALESCE(NULLIF(cf_continent, ''), '') AS continent,
-    session,
-    COALESCE(NULLIF(workspace, ''), '') AS workspace,
-    COALESCE(NULLIF(api_key, ''), '') AS api_key,
+    model_requested AS raw_model,
+    ${statModelSql("model_requested", "route_model")} AS model,
+    COALESCE(NULLIF(route_model, ''), '') AS provider_model,
+    COALESCE(NULLIF(provider_id, ''), '') AS raw_provider,
+    UPPER(COALESCE(NULLIF(country, ''), 'ZZ')) AS country,
+    COALESCE(NULLIF(continent, ''), '') AS continent,
+    session_id AS session,
+    COALESCE(NULLIF(workspace_id, ''), '') AS workspace,
+    COALESCE(NULLIF(service_api_key_id, ''), '') AS api_key,
    COALESCE(NULLIF(user_id, ''), '') AS user_id,
-    status,
-    duration AS duration_ms,
-    time_to_first_byte AS ttfb_ms,
-    timestamp_first_byte,
-    timestamp_last_byte,
+    outcome,
+    duration_ms,
+    time_to_first_token_ms AS ttfb_ms,
+    CASE
+      WHEN first_token_at IS NULL OR last_token_at IS NULL THEN null
+      ELSE date_part('epoch', last_token_at) - date_part('epoch', first_token_at)
+    END AS output_seconds,
    tokens_input,
    tokens_output,
    tokens_reasoning,
    tokens_cache_read,
-    tokens_cache_write_5m,
-    tokens_cache_write_1h,
-    cost_input_microcents,
-    cost_output_microcents,
-    cost_total_microcents,
-    cost_input,
-    cost_output,
-    cost_total,
-    source
+    tokens_cache_write,
+    cost_input AS cost_input_microcents,
+    cost_output AS cost_output_microcents,
+    cost_total AS cost_total_microcents
  FROM ${sourceTable}
-  WHERE event_type = 'completions'
-    AND model IS NOT NULL
-    AND model <> ''
-    AND source = 'lite'
-    AND event_date >= ${periodStartDateValue}
-    AND event_date <= ${periodEndDateValue}
-    AND event_timestamp >= ${periodStartValue}
-    AND event_timestamp < ${periodEndValue}
+  WHERE event_type = 'generation.completed'
+    AND source IN ('inference', 'inference-legacy')
+    AND (
+      (source = 'inference-legacy' AND started_at < ${sqlString(LIVE_SOURCE_START)})
+      OR (source = 'inference' AND started_at >= ${sqlString(LIVE_SOURCE_START)})
+    )
+    AND product = 'go'
+    AND model_requested IS NOT NULL
+    AND model_requested <> ''
+    AND __ingest_ts >= ${periodStartValue}
+    AND __ingest_ts < ${ingestEndValue}
+    AND started_at >= ${periodStartValue}
+    AND started_at < ${periodEndValue}
 ), filtered AS (
  SELECT
-    event_time,
-    CASE
-      WHEN source = 'lite' THEN 'Go'
-      WHEN raw_model IN ('gpt-5-nano', 'grok-code', 'big-pickle') OR regexp_like(raw_model, '-free(:global)?$') THEN 'Free'
-      ELSE 'Paid'
-    END AS tier,
+    'Go' AS tier,
    ${statProviderSql("model", "provider_model", "raw_provider")} AS provider,
    provider_model,
    model,
@@ -104,63 +155,39 @@ WITH normalized AS (
    continent,
    session,
    COALESCE(NULLIF(user_id, ''), NULLIF(workspace, ''), NULLIF(api_key, '')) AS user_key,
-    status,
+    outcome,
    duration_ms,
    ttfb_ms,
    CASE
-      WHEN timestamp_last_byte - timestamp_first_byte < 100 THEN null
-      ELSE CAST(tokens_output AS double) / (timestamp_last_byte - timestamp_first_byte) * 1000
+      WHEN output_seconds < 0.1 THEN null
+      ELSE CAST(tokens_output AS double) / output_seconds
    END AS output_tps,
    tokens_input,
    tokens_output,
    tokens_reasoning,
    tokens_cache_read,
-    COALESCE(tokens_cache_read, 0) + COALESCE(tokens_cache_write_5m, 0) + COALESCE(tokens_cache_write_1h, 0) + COALESCE(tokens_input, 0) + COALESCE(tokens_output, 0) AS tokens_total,
-    COALESCE(cost_input_microcents, cost_input * 1000000) AS cost_input_microcents,
-    COALESCE(cost_output_microcents, cost_output * 1000000) AS cost_output_microcents,
-    COALESCE(cost_total_microcents, cost_total * 1000000) AS cost_total_microcents
+    COALESCE(tokens_cache_read, 0) + COALESCE(tokens_cache_write, 0) + COALESCE(tokens_input, 0) + COALESCE(tokens_output, 0) AS tokens_total,
+    cost_input_microcents,
+    cost_output_microcents,
+    cost_total_microcents
  FROM normalized
  WHERE lower(model) NOT IN (${[...EXCLUDED_MODELS].map(sqlString).join(", ")})
-), periods AS (
-  SELECT
-    concat(CAST(year_of_week(event_time) AS varchar), '-W', lpad(CAST(week(event_time) AS varchar), 2, '0')) AS week_key,
-    substr(to_iso8601(date_trunc('day', event_time)), 1, 10) AS day_key,
-    *
-  FROM filtered
 )
 SELECT
-  CASE WHEN grouping(week_key) = 0 THEN 'week' ELSE 'day' END AS grain,
-  COALESCE(week_key, day_key) AS period_key,
-  ${sqlString(Resource.StatsSyncConfig.dataset)} AS dataset,
-  CASE
-    WHEN grouping(country) = 0 AND grouping(model) = 0 THEN 'geo_model'
-    WHEN grouping(country) = 0 THEN 'geo'
-    WHEN grouping(model) = 0 THEN 'model'
-    ELSE 'provider'
-  END AS dimension,
-  tier,
-  CASE WHEN grouping(provider) = 0 THEN provider ELSE 'all' END AS provider,
-  CASE WHEN grouping(model) = 0 THEN model WHEN grouping(country) = 0 THEN 'all' END AS model,
-  CASE WHEN grouping(model) = 0 AND grouping(country) = 1 THEN COALESCE(MAX(NULLIF(provider_model, '')), '') END AS provider_model,
-  CASE WHEN grouping(country) = 0 THEN country END AS country,
-  CASE WHEN grouping(country) = 0 THEN COALESCE(MAX(NULLIF(continent, '')), '') END AS continent,
+  ${sqlString(period.grain)} AS grain,
+  ${sqlString(period.key)} AS period_key,
+  ${sqlString(source.dataset)} AS dataset,
+  ${dimensions},
  ${aggregateColumns}
-FROM periods
+FROM filtered
 GROUP BY GROUPING SETS (
-  (week_key, tier, provider, model),
-  (week_key, tier, provider),
-  (week_key, tier, country),
-  (week_key, tier, provider, model, country),
-  (day_key, tier, provider, model),
-  (day_key, tier, provider),
-  (day_key, tier, country),
-  (day_key, tier, provider, model, country)
+  ${groupingSets}
 )
-ORDER BY grain, period_key, total_tokens DESC
+LIMIT 10000
 `
 }

-export function toModelAggregate(data: AthenaData): ModelStatAggregate[] {
+export function toModelAggregate(data: R2SqlData): ModelStatAggregate[] {
  const model = statModel(data.model, data.provider_model)
  const provider = statProvider(model, data.provider_model, data.provider)
  if (!provider) return []
@@ -170,13 +197,13 @@ export function toModelAggregate(data: AthenaData): ModelStatAggregate[] {
  ])
 }

-export function toProviderAggregate(data: AthenaData): ProviderStatAggregate[] {
+export function toProviderAggregate(data: R2SqlData): ProviderStatAggregate[] {
  return toStatBaseAggregate(data).flatMap((base) => [
    { ...base, provider: statProvider(data.model, data.provider_model, data.provider) || "unknown" },
  ])
 }

-export function toGeoAggregate(data: AthenaData): GeoStatAggregate[] {
+export function toGeoAggregate(data: R2SqlData): GeoStatAggregate[] {
  return toStatBaseAggregate(data).flatMap((base) => [
    {
      ...base,
@@ -188,7 +215,7 @@ export function toGeoAggregate(data: AthenaData): GeoStatAggregate[] {
  ])
 }

-function toStatBaseAggregate(data: AthenaData): StatBaseAggregate[] {
+function toStatBaseAggregate(data: R2SqlData): StatBaseAggregate[] {
  const grain = data.grain === "day" || data.grain === "week" ? data.grain : undefined
  if (!grain || !data.period_key) return []

@@ -223,21 +250,21 @@ function toStatBaseAggregate(data: AthenaData): StatBaseAggregate[] {
  ]
 }

-function integer(data: AthenaData, key: string) {
+function integer(data: R2SqlData, key: string) {
  return Math.round(number(data, key))
 }

-function nullableNumber(data: AthenaData, key: string) {
+function nullableNumber(data: R2SqlData, key: string) {
  if (data[key] === undefined || data[key] === "") return null
  return Number(number(data, key).toFixed(2))
 }

-function nullableInteger(data: AthenaData, key: string) {
+function nullableInteger(data: R2SqlData, key: string) {
  if (data[key] === undefined || data[key] === "") return null
  return Math.round(number(data, key))
 }

-function number(data: AthenaData, key: string) {
+function number(data: R2SqlData, key: string) {
  const value = Number(data[key])
  return Number.isFinite(value) ? value : 0
 }
@@ -250,6 +277,21 @@ function sqlString(value: string) {
  return `'${value.replace(/'/g, "''")}'`
 }

+function statPeriods(grain: "day" | "week", periodStart: Date, periodEnd: Date) {
+  const interval = grain === "day" ? DAY_MS : WEEK_MS
+  const first = grain === "day" ? startOfUtcDay(periodStart) : startOfIsoWeek(periodStart)
+  const count = Math.max(0, Math.ceil((periodEnd.getTime() - first.getTime()) / interval))
+  return Array.from({ length: count }, (_, index) => {
+    const start = new Date(first.getTime() + index * interval)
+    return {
+      grain,
+      key: periodKeyFor(grain, start),
+      start,
+      end: new Date(Math.min(start.getTime() + interval, periodEnd.getTime())),
+    }
+  })
+}
+
 function statModelSql(model: string, providerModel: string) {
  return `COALESCE(NULLIF(regexp_replace(CASE
      WHEN lower(${model}) = 'big-pickle' THEN NULLIF(${providerModel}, '')
@@ -0,0 +1,105 @@
+import { Context, Effect, Layer, Schema } from "effect"
+import { Resource } from "sst/resource"
+
+const R2_SQL_MAX_ROWS = 10_000
+const R2SqlValue = Schema.Union([Schema.String, Schema.Number, Schema.Boolean, Schema.Null])
+const R2SqlResponse = Schema.Struct({
+  success: Schema.Boolean,
+  result: Schema.optional(
+    Schema.NullOr(
+      Schema.Struct({
+        request_id: Schema.String,
+        rows: Schema.Array(Schema.Record(Schema.String, R2SqlValue)),
+      }),
+    ),
+  ),
+  errors: Schema.Array(Schema.Unknown),
+})
+const decodeResponse = Schema.decodeUnknownEffect(Schema.fromJsonString(R2SqlResponse))
+
+export type R2SqlData = Record<string, string>
+
+export class R2SqlQueryError extends Error {
+  readonly _tag = "R2SqlQueryError"
+  readonly requestId?: string
+  readonly status?: number
+
+  constructor(input: { message: string; requestId?: string; status?: number; cause?: unknown }) {
+    super(input.message, { cause: input.cause })
+    this.name = "R2SqlQueryError"
+    this.requestId = input.requestId
+    this.status = input.status
+  }
+}
+
+export declare namespace R2Sql {
+  export interface Service {
+    readonly query: (query: string) => Effect.Effect<R2SqlData[], R2SqlQueryError>
+  }
+}
+
+export class R2Sql extends Context.Service<R2Sql, R2Sql.Service>()("@opencode/stats/R2Sql") {
+  static readonly layer: Layer.Layer<R2Sql> = Layer.succeed(
+    R2Sql,
+    R2Sql.of({
+      query: Effect.fn("R2Sql.query")(function* (query: string) {
+        const response = yield* Effect.tryPromise({
+          try: () =>
+            Bun.fetch(
+              `https://api.sql.cloudflarestorage.com/api/v1/accounts/${Resource.R2Sql.accountId}/r2-sql/query/${Resource.R2Sql.bucket}`,
+              {
+                method: "POST",
+                headers: {
+                  Authorization: `Bearer ${Resource.R2SqlAuthToken.value}`,
+                  "Content-Type": "application/json",
+                },
+                body: JSON.stringify({ query }),
+              },
+            ),
+          catch: (cause) => new R2SqlQueryError({ message: "Failed to run R2 SQL stats query", cause }),
+        })
+        const body = yield* Effect.tryPromise({
+          try: () => response.text(),
+          catch: (cause) =>
+            new R2SqlQueryError({ message: "Failed to read R2 SQL stats response", status: response.status, cause }),
+        })
+        const decoded = yield* decodeResponse(body).pipe(
+          Effect.mapError(
+            (cause) =>
+              new R2SqlQueryError({
+                message: "R2 SQL returned an invalid stats response",
+                status: response.status,
+                cause,
+              }),
+          ),
+        )
+        if (!response.ok || !decoded.success || !decoded.result)
+          return yield* Effect.fail(
+            new R2SqlQueryError({
+              message: `R2 SQL stats query failed: ${JSON.stringify(decoded.errors)}`,
+              requestId: decoded.result?.request_id,
+              status: response.status,
+            }),
+          )
+
+        // R2 SQL has no OFFSET support and caps LIMIT at 10,000. Each stats
+        // query is scoped to one day or week, and reaching the cap is treated as
+        // an error so a newly high-cardinality period can never be truncated.
+        if (decoded.result.rows.length >= R2_SQL_MAX_ROWS)
+          return yield* Effect.fail(
+            new R2SqlQueryError({
+              message: `R2 SQL stats query reached the ${R2_SQL_MAX_ROWS} row limit`,
+              requestId: decoded.result.request_id,
+              status: response.status,
+            }),
+          )
+
+        return decoded.result.rows.map((row) =>
+          Object.fromEntries(
+            Object.entries(row).flatMap(([key, value]) => (value === null ? [] : [[key, String(value)]])),
+          ),
+        )
+      }),
+    }),
+  )
+}
@@ -11,6 +11,17 @@ declare module "sst/resource" {
      type: "sst.sst.Linkable"
      workgroup: string
    }
+    R2Sql: {
+      accountId: string
+      bucket: string
+      namespace: string
+      table: string
+      type: "sst.sst.Linkable"
+    }
+    R2SqlAuthToken: {
+      type: "sst.sst.Secret"
+      value: string
+    }
    StatsSyncConfig: {
      dataset: string
      type: "sst.sst.Linkable"
@@ -1,12 +1,12 @@
 import { DateTime, Effect } from "effect"
 import { Resource } from "sst/resource"
-import { Athena, AthenaQueryError, AthenaQueryTimeoutError } from "./athena"
 import { DatabaseError } from "./database"
 import { GeoStatRepo, rowsFromAggregates as geoRowsFromAggregates } from "./domain/geo"
-import { buildStatsQuery, toGeoAggregate, toModelAggregate, toProviderAggregate } from "./domain/inference"
+import { buildStatsQueries, toGeoAggregate, toModelAggregate, toProviderAggregate } from "./domain/inference"
 import { ModelStatRepo, rowsFromAggregates as modelRowsFromAggregates } from "./domain/model"
 import { ProviderStatRepo, rowsFromAggregates as providerRowsFromAggregates } from "./domain/provider"
 import { startOfIsoWeek } from "./domain/stat"
+import { R2Sql, R2SqlQueryError } from "./r2-sql"

 const DATALAKE_INGESTION_LAG_MS = 5 * 60_000
 const STATS_DATA_START_MS = new Date("2026-05-28T00:00:00.000Z").getTime()
@@ -18,23 +18,25 @@ const DISPLAY_WINDOW_MS = 56 * 86_400_000
 const INCREMENTAL_LOOKBACK_MS = 2 * 3_600_000

 export type SyncStatsResult = { ok: true; rows: number; startedAt: string; periodStart: string; periodEnd: string }
-export type SyncStatsError = AthenaQueryError | AthenaQueryTimeoutError | DatabaseError
+export type SyncStatsError = R2SqlQueryError | DatabaseError

 export const syncStats: (options?: {
  full?: boolean
-}) => Effect.Effect<SyncStatsResult, SyncStatsError, Athena | ModelStatRepo | ProviderStatRepo | GeoStatRepo> =
+}) => Effect.Effect<SyncStatsResult, SyncStatsError, R2Sql | ModelStatRepo | ProviderStatRepo | GeoStatRepo> =
  Effect.fn("StatSync.sync")(function* (options?: { full?: boolean }) {
    const startedAt = yield* DateTime.nowAsDate
    const periodEnd = new Date(Math.floor((startedAt.getTime() - DATALAKE_INGESTION_LAG_MS) / 60_000) * 60_000)
    const periodStart = options?.full ? fullPeriodStart(periodEnd) : incrementalPeriodStart(periodEnd)
-    const athena = yield* Athena
+    const r2Sql = yield* R2Sql
    const modelStats = yield* ModelStatRepo
    const providerStats = yield* ProviderStatRepo
    const geoStats = yield* GeoStatRepo

    yield* logRuntimeCheck()

-    const rows = yield* athena.query(buildStatsQuery(periodStart, periodEnd))
+    const rows = yield* Effect.forEach(buildStatsQueries(periodStart, periodEnd), r2Sql.query, {
+      concurrency: 4,
+    }).pipe(Effect.map((batches) => batches.flat()))
    const modelRows = modelRowsFromAggregates(rows.filter((row) => row.dimension === "model").flatMap(toModelAggregate))
    const providerRows = providerRowsFromAggregates(
      rows.filter((row) => row.dimension === "provider").flatMap(toProviderAggregate),
@@ -77,7 +79,7 @@ export const syncStats: (options?: {
    }
  })

-// May 27 was partial, so keep Athena stats anchored at the first complete day.
+// May 27 was partial, so keep stats anchored at the first complete day.
 function fullPeriodStart(periodEnd: Date) {
  return new Date(
    Math.max(
@@ -99,13 +101,12 @@ function incrementalPeriodStart(periodEnd: Date) {

 function logRuntimeCheck() {
  return Effect.logInfo(
-    `athena stats runtime check ${JSON.stringify({
-      catalog: Resource.InferenceEvent.catalog,
-      database: Resource.InferenceEvent.database,
+    `r2 sql stats runtime check ${JSON.stringify({
+      accountId: Resource.R2Sql.accountId,
+      bucket: Resource.R2Sql.bucket,
      dataset: Resource.StatsSyncConfig.dataset,
-      table: Resource.InferenceEvent.table,
-      workgroup: Resource.InferenceEvent.workgroup,
-      region: Resource.InferenceEvent.region,
+      namespace: Resource.R2Sql.namespace,
+      table: Resource.R2Sql.table,
      stage: Resource.App.stage,
    })}`,
  )
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/stats-server",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 import * as NodeRuntime from "@effect/platform-node/NodeRuntime"
-import { Athena } from "@opencode-ai/stats-core/athena"
 import { ModelStatRepo } from "@opencode-ai/stats-core/domain/model"
+import { R2Sql } from "@opencode-ai/stats-core/r2-sql"
 import { layer as statsLayer } from "@opencode-ai/stats-core/runtime"
 import { syncStats } from "@opencode-ai/stats-core/stat-sync"
 import { Cause, Duration, Effect, Layer, Schedule } from "effect"
@@ -8,7 +8,7 @@ import { Cause, Duration, Effect, Layer, Schedule } from "effect"
 const SYNC_INTERVAL = "1 hour"
 const SYNC_INTERVAL_MS = 3_600_000

-const runtimeLayer = Layer.mergeAll(statsLayer, Athena.layer)
+const runtimeLayer = Layer.mergeAll(statsLayer, R2Sql.layer)

 const daemon = Effect.gen(function* () {
  yield* Effect.logInfo("stats sync daemon started")
@@ -19,9 +19,19 @@ const daemon = Effect.gen(function* () {
  let lastFullDay = ""
  const pass = Effect.gen(function* () {
    const today = new Date().toISOString().slice(0, 10)
-    const full = lastFullDay !== today
-    yield* syncStats({ full })
-    if (full) lastFullDay = today
+    if (lastFullDay !== today) {
+      const completed = yield* syncStats({ full: true }).pipe(
+        Effect.as(true),
+        Effect.catchCause((cause) =>
+          Effect.logWarning(`full stats sync failed; falling back to incremental sync ${Cause.pretty(cause)}`).pipe(
+            Effect.as(false),
+          ),
+        ),
+      )
+      lastFullDay = today
+      if (completed) return
+    }
+    yield* syncStats({ full: false })
  }).pipe(
    Effect.catchCause((cause) =>
      Effect.logWarning(`stats sync failed ${JSON.stringify({ cause: Cause.pretty(cause) })}`),
@@ -30,9 +40,9 @@ const daemon = Effect.gen(function* () {
  yield* pass.pipe(Effect.repeat(Schedule.fixed(SYNC_INTERVAL)))
 }).pipe(Effect.forkScoped)

-// A restarted daemon must not immediately re-run the expensive Athena pass; resume
-// the hourly cadence from the last completed sync instead. This caps the Athena
-// spend of a crash loop at one pass per interval.
+// A restarted daemon must not immediately re-run the R2 SQL pass; resume the
+// hourly cadence from the last completed sync instead. This caps the query spend
+// of a crash loop at one pass per interval.
 const initialDelay = Effect.fnUntraced(function* () {
  const modelStats = yield* ModelStatRepo
  const lastSynced = yield* modelStats.lastSyncedAt().pipe(Effect.catchCause(() => Effect.succeed(null)))
@@ -1,7 +1,7 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
  "name": "@opencode-ai/tui",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "private": true,
  "type": "module",
  "license": "MIT",
@@ -1,6 +1,6 @@
 {
  "name": "@opencode-ai/ui",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "type": "module",
  "license": "MIT",
  "repository": {
@@ -25,7 +25,7 @@
      "border-weak-base": "#DBDBDB",
      "border-weaker-base": "#E8E8E8",
      "icon-base": "#8F8F8F",
-      "icon-weak-base": "C7C7C7",
+      "icon-weak-base": "#C7C7C7",
      "surface-raised-base": "#F3F3F3",
      "surface-raised-base-hover": "#EDEDED",
      "surface-base": "#F8F8F8",
@@ -195,9 +195,9 @@

 /* Loading */
 [data-component="button-v2"][data-variant="loading"] {
-  background: #f2f2f2;
+  background: var(--v2-background-bg-layer-02);
  color: var(--v2-text-text-base);
-  box-shadow: inset 0 0 0 0.5px rgba(0, 0, 0, 0.08);
+  box-shadow: inset 0 0 0 0.5px var(--v2-border-border-muted);
  cursor: default;
  pointer-events: none;
 }
@@ -2,7 +2,7 @@
  "name": "@opencode-ai/web",
  "type": "module",
  "license": "MIT",
-  "version": "1.18.14",
+  "version": "1.18.18",
  "scripts": {
    "dev": "astro dev",
    "dev:remote": "VITE_API_URL=https://api.opencode.ai astro dev",
@@ -185,7 +185,7 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال

 | Model             | Model ID          | Endpoint                                         | AI SDK Package              |
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
-| Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| 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-5.2           | 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` |
@@ -53,69 +53,69 @@ OpenCode Zen هي بوابة AI تتيح لك الوصول إلى هذه الن

 يمكنك أيضا الوصول إلى نماذجنا عبر نقاط نهاية API التالية.

-| النموذج                | معرّف النموذج          | نقطة النهاية                                              | حزمة AI SDK                 |
-| ---------------------- | ---------------------- | --------------------------------------------------------- | --------------------------- |
-| GPT 5.6 Sol            | gpt-5.6-sol            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Terra          | gpt-5.6-terra          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Luna           | gpt-5.6-luna           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5                | gpt-5.5                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5 Pro            | gpt-5.5-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4                | gpt-5.4                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Pro            | gpt-5.4-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Mini           | gpt-5.4-mini           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Nano           | gpt-5.4-nano           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex          | gpt-5.3-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex Spark    | gpt-5.3-codex-spark    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2                | gpt-5.2                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2 Codex          | gpt-5.2-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1                | gpt-5.1                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex          | gpt-5.1-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Max      | gpt-5.1-codex-max      | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Mini     | gpt-5.1-codex-mini     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5                  | gpt-5                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Codex            | gpt-5-codex            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Nano             | gpt-5-nano             | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Claude Fable 5         | claude-fable-5         | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 5          | claude-opus-5          | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.8        | claude-opus-4-8        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.7        | claude-opus-4-7        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.6        | claude-opus-4-6        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.5        | claude-opus-4-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 5        | claude-sonnet-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.6      | claude-sonnet-4-6      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.5      | claude-sonnet-4-5      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Haiku 4.5       | claude-haiku-4-5       | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Gemini 3.6 Flash       | gemini-3.6-flash       | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash       | gemini-3.5-flash       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash Lite  | gemini-3.5-flash-lite  | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
-| Gemini 3.1 Pro         | gemini-3.1-pro         | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
-| Gemini 3 Flash         | gemini-3-flash         | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
-| Grok 4.5               | grok-4.5               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Grok Build 0.1         | grok-build-0.1         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Qwen3.7 Max            | qwen3.7-max            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus           | qwen3.7-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus           | qwen3.6-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.5 Plus           | qwen3.5-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| DeepSeek V4 Pro        | deepseek-v4-pro        | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash      | deepseek-v4-flash      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M3             | minimax-m3             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.7           | minimax-m2.7           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.5           | minimax-m2.5           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.2                | glm-5.2                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.1                | glm-5.1                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5                  | glm-5                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.5              | kimi-k2.5              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.6              | kimi-k2.6              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code         | kimi-k2.7-code         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K3                | kimi-k3                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Big Pickle             | big-pickle             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5 Free         | mimo-v2.5-free         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Laguna S 2.1 Free      | laguna-s-2.1-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Ling-3.0-flash Free    | ling-3.0-flash-free    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| LongCat-2.0 Free       | longcat-2.0-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| North Mini Code Free   | north-mini-code-free   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Nemotron 3 Ultra Free  | nemotron-3-ultra-free  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash Free | deepseek-v4-flash-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| النموذج                     | معرّف النموذج               | نقطة النهاية                                              | حزمة AI SDK                 |
+| --------------------------- | --------------------------- | --------------------------------------------------------- | --------------------------- |
+| GPT 5.6 Sol                 | gpt-5.6-sol                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Terra               | gpt-5.6-terra               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                | gpt-5.6-luna                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5                     | gpt-5.5                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5 Pro                 | gpt-5.5-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4                     | gpt-5.4                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Pro                 | gpt-5.4-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Mini                | gpt-5.4-mini                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Nano                | gpt-5.4-nano                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex               | gpt-5.3-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex Spark         | gpt-5.3-codex-spark         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2                     | gpt-5.2                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2 Codex               | gpt-5.2-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1                     | gpt-5.1                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex               | gpt-5.1-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Max           | gpt-5.1-codex-max           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Mini          | gpt-5.1-codex-mini          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5                       | gpt-5                       | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Codex                 | gpt-5-codex                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Nano                  | gpt-5-nano                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Claude Fable 5              | claude-fable-5              | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 5               | claude-opus-5               | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.8             | claude-opus-4-8             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.7             | claude-opus-4-7             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.6             | claude-opus-4-6             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.5             | claude-opus-4-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 5             | claude-sonnet-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.6           | claude-sonnet-4-6           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.5           | claude-sonnet-4-5           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Haiku 4.5            | claude-haiku-4-5            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Gemini 3.6 Flash            | gemini-3.6-flash            | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash            | gemini-3.5-flash            | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash Lite       | gemini-3.5-flash-lite       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
+| Gemini 3.1 Pro              | gemini-3.1-pro              | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
+| Gemini 3 Flash              | gemini-3-flash              | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
+| Grok 4.6                    | grok-4.6                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok 4.5                    | grok-4.5                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok Build 0.1              | grok-build-0.1              | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Qwen3.7 Max                 | qwen3.7-max                 | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                | qwen3.7-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                | qwen3.6-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.5 Plus                | qwen3.5-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| DeepSeek V4 Pro             | deepseek-v4-pro             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash           | deepseek-v4-flash           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M3                  | minimax-m3                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.7                | minimax-m2.7                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.5                | minimax-m2.5                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.2                     | glm-5.2                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.1                     | glm-5.1                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5                       | glm-5                       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.5                   | kimi-k2.5                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                   | kimi-k2.6                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code              | kimi-k2.7-code              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K3                     | kimi-k3                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Big Pickle                  | big-pickle                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5 Free              | mimo-v2.5-free              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Hy3 Free                    | hy3-free                    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Laguna S 2.1 Free           | laguna-s-2.1-free           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3 Ultra Free       | nemotron-3-ultra-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3.5 Lightning Free | nemotron-3.5-lightning-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Free      | deepseek-v4-flash-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |

 يستخدم [معرّف النموذج](/docs/config/#models) في إعدادات OpenCode الصيغة `opencode/<model-id>`. على سبيل المثال، بالنسبة إلى GPT 5.5، ستستخدم `opencode/gpt-5.5` في إعداداتك.

@@ -140,11 +140,10 @@ https://opencode.ai/zen/v1/models
 | Big Pickle                        | Free    | Free    | Free            | -               |
 | DeepSeek V4 Flash Free            | Free    | Free    | Free            | -               |
 | MiMo-V2.5 Free                    | Free    | Free    | Free            | -               |
+| Hy3 Free                          | Free    | Free    | Free            | -               |
 | Laguna S 2.1 Free                 | Free    | Free    | Free            | -               |
-| Ling-3.0-flash Free               | Free    | Free    | Free            | -               |
-| LongCat-2.0 Free                  | Free    | Free    | Free            | -               |
-| North Mini Code Free              | Free    | Free    | Free            | -               |
 | Nemotron 3 Ultra Free             | Free    | Free    | Free            | -               |
+| Nemotron 3.5 Lightning Free       | Free    | Free    | Free            | -               |
 | MiniMax M3                        | $0.30   | $1.20   | $0.06           | -               |
 | MiniMax M2.7                      | $0.30   | $1.20   | $0.06           | -               |
 | MiniMax M2.5                      | $0.30   | $1.20   | $0.06           | -               |
@@ -178,6 +177,8 @@ https://opencode.ai/zen/v1/models
 | Gemini 3.1 Pro (≤ 200K tokens)    | $2.00   | $12.00  | $0.20           | -               |
 | Gemini 3.1 Pro (> 200K tokens)    | $4.00   | $18.00  | $0.40           | -               |
 | Gemini 3 Flash                    | $0.50   | $3.00   | $0.05           | -               |
+| Grok 4.6 (≤ 200K tokens)          | $2.00   | $6.00   | $0.50           | -               |
+| Grok 4.6 (> 200K tokens)          | $4.00   | $12.00  | $1.00           | -               |
 | Grok 4.5 (≤ 200K tokens)          | $2.00   | $6.00   | $0.30           | -               |
 | Grok 4.5 (> 200K tokens)          | $4.00   | $12.00  | $0.60           | -               |
 | Grok Build 0.1                    | $1.00   | $2.00   | $0.20           | -               |
@@ -217,11 +218,10 @@ https://opencode.ai/zen/v1/models

 - DeepSeek V4 Flash Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
 - MiMo-V2.5 Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
+- Hy3 Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
 - Laguna S 2.1 Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
- Ling-3.0-flash Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
- LongCat-2.0 Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
- North Mini Code Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
 - Nemotron 3 Ultra Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
+- Nemotron 3.5 Lightning Free متاح على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.
 - Big Pickle نموذج خفي ومتاح مجانا على OpenCode لفترة محدودة. يستخدم الفريق هذه الفترة لجمع الملاحظات وتحسين النموذج.

 <a href={email}>تواصل معنا</a> إذا كانت لديك أي أسئلة.
@@ -276,11 +276,10 @@ https://opencode.ai/zen/v1/models
 - Big Pickle: خلال فترته المجانية، قد تُستخدم البيانات المجمعة لتحسين النموذج.
 - DeepSeek V4 Flash Free: خلال فترته المجانية، قد تُستخدم البيانات المجمعة لتحسين النموذج.
 - MiMo-V2.5 Free: خلال فترته المجانية، قد تُستخدم البيانات المجمعة لتحسين النموذج.
+- Hy3 Free: خلال فترته المجانية، قد تُستخدم البيانات المجمعة لتحسين النموذج.
 - Laguna S 2.1 Free: خلال فترته المجانية، قد تُستخدم البيانات المجمعة لتحسين النموذج.
- Ling-3.0-flash Free: خلال فترته المجانية، قد تُستخدم البيانات المجمعة لتحسين النموذج.
- LongCat-2.0 Free: خلال فترته المجانية، قد تُستخدم البيانات المجمعة لتحسين النموذج.
- North Mini Code Free: خلال فترته المجانية، قد يُحتفَظ بالبيانات المُجمَّعة وتُستخدم لتحسين النموذج. لا تُرسل بيانات شخصية أو سرية. راجع [شروط الاستخدام](https://cohere.com/terms-of-use) و[سياسة الخصوصية](https://cohere.com/privacy).
 - Nemotron 3 Ultra Free (نقاط نهاية NVIDIA المجانية): للاستخدام التجريبي فقط — لا ترسل بيانات شخصية أو سرية. يُسجَّل استخدامك لأغراض أمنية ولتحسين منتجات وخدمات NVIDIA. بيانات الجلسة المُسجَّلة لأغراض التحسين غير مرتبطة بهويتك أو بأي مُعرِّف دائم. لمزيد من المعلومات حول ممارسات معالجة البيانات لدينا، راجع [سياسة الخصوصية](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). بتفاعلك مع نقطة النهاية هذه، فإنك توافق على جمعنا لهذه المعلومات وتسجيلها واستخدامها وعلى [شروط خدمة النسخة التجريبية من واجهة NVIDIA API](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).
+- Nemotron 3.5 Lightning Free (نقاط نهاية NVIDIA المجانية): للاستخدام التجريبي فقط — لا ترسل بيانات شخصية أو سرية. يُسجَّل استخدامك لأغراض أمنية ولتحسين منتجات وخدمات NVIDIA. بيانات الجلسة المُسجَّلة لأغراض التحسين غير مرتبطة بهويتك أو بأي مُعرِّف دائم. لمزيد من المعلومات حول ممارسات معالجة البيانات لدينا، راجع [سياسة الخصوصية](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). بتفاعلك مع نقطة النهاية هذه، فإنك توافق على جمعنا لهذه المعلومات وتسجيلها واستخدامها وعلى [شروط خدمة النسخة التجريبية من واجهة NVIDIA API](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).
 - OpenAI APIs: يتم الاحتفاظ بالطلبات لمدة 30 يوما وفقا لـ [OpenAI's Data Policies](https://platform.openai.com/docs/guides/your-data).
 - Anthropic APIs: يتم الاحتفاظ بالطلبات لمدة 30 يوما وفقا لـ [Anthropic's Data Policies](https://docs.anthropic.com/en/docs/claude-code/data-usage).

@@ -197,7 +197,7 @@ 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/chat/completions` | `@ai-sdk/openai-compatible` |
+| 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-5.2           | 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` |
@@ -58,69 +58,69 @@ Naplata se vrši po zahtjevu i možete dodavati kredit na svoj račun.

 Našim modelima možete pristupiti i preko sljedećih API endpointa.

-| Model                  | Model ID               | Endpoint                                                  | AI SDK Package              |
-| ---------------------- | ---------------------- | --------------------------------------------------------- | --------------------------- |
-| GPT 5.6 Sol            | gpt-5.6-sol            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Terra          | gpt-5.6-terra          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Luna           | gpt-5.6-luna           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5                | gpt-5.5                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5 Pro            | gpt-5.5-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4                | gpt-5.4                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Pro            | gpt-5.4-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Mini           | gpt-5.4-mini           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Nano           | gpt-5.4-nano           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex          | gpt-5.3-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex Spark    | gpt-5.3-codex-spark    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2                | gpt-5.2                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2 Codex          | gpt-5.2-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1                | gpt-5.1                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex          | gpt-5.1-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Max      | gpt-5.1-codex-max      | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Mini     | gpt-5.1-codex-mini     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5                  | gpt-5                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Codex            | gpt-5-codex            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Nano             | gpt-5-nano             | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Claude Fable 5         | claude-fable-5         | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 5          | claude-opus-5          | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.8        | claude-opus-4-8        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.7        | claude-opus-4-7        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.6        | claude-opus-4-6        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.5        | claude-opus-4-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 5        | claude-sonnet-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.6      | claude-sonnet-4-6      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.5      | claude-sonnet-4-5      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Haiku 4.5       | claude-haiku-4-5       | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Gemini 3.6 Flash       | gemini-3.6-flash       | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash       | gemini-3.5-flash       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash Lite  | gemini-3.5-flash-lite  | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
-| Gemini 3.1 Pro         | gemini-3.1-pro         | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
-| Gemini 3 Flash         | gemini-3-flash         | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
-| Grok 4.5               | grok-4.5               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Grok Build 0.1         | grok-build-0.1         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Qwen3.7 Max            | qwen3.7-max            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus           | qwen3.7-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus           | qwen3.6-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.5 Plus           | qwen3.5-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| DeepSeek V4 Pro        | deepseek-v4-pro        | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash      | deepseek-v4-flash      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M3             | minimax-m3             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.7           | minimax-m2.7           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.5           | minimax-m2.5           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.2                | glm-5.2                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.1                | glm-5.1                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5                  | glm-5                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.5              | kimi-k2.5              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.6              | kimi-k2.6              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code         | kimi-k2.7-code         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K3                | kimi-k3                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Big Pickle             | big-pickle             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5 Free         | mimo-v2.5-free         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Laguna S 2.1 Free      | laguna-s-2.1-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Ling-3.0-flash Free    | ling-3.0-flash-free    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| LongCat-2.0 Free       | longcat-2.0-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| North Mini Code Free   | north-mini-code-free   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Nemotron 3 Ultra Free  | nemotron-3-ultra-free  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash Free | deepseek-v4-flash-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Model                       | Model ID                    | Endpoint                                                  | AI SDK Package              |
+| --------------------------- | --------------------------- | --------------------------------------------------------- | --------------------------- |
+| GPT 5.6 Sol                 | gpt-5.6-sol                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Terra               | gpt-5.6-terra               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                | gpt-5.6-luna                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5                     | gpt-5.5                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5 Pro                 | gpt-5.5-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4                     | gpt-5.4                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Pro                 | gpt-5.4-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Mini                | gpt-5.4-mini                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Nano                | gpt-5.4-nano                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex               | gpt-5.3-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex Spark         | gpt-5.3-codex-spark         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2                     | gpt-5.2                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2 Codex               | gpt-5.2-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1                     | gpt-5.1                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex               | gpt-5.1-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Max           | gpt-5.1-codex-max           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Mini          | gpt-5.1-codex-mini          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5                       | gpt-5                       | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Codex                 | gpt-5-codex                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Nano                  | gpt-5-nano                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Claude Fable 5              | claude-fable-5              | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 5               | claude-opus-5               | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.8             | claude-opus-4-8             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.7             | claude-opus-4-7             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.6             | claude-opus-4-6             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.5             | claude-opus-4-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 5             | claude-sonnet-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.6           | claude-sonnet-4-6           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.5           | claude-sonnet-4-5           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Haiku 4.5            | claude-haiku-4-5            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Gemini 3.6 Flash            | gemini-3.6-flash            | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash            | gemini-3.5-flash            | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash Lite       | gemini-3.5-flash-lite       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
+| Gemini 3.1 Pro              | gemini-3.1-pro              | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
+| Gemini 3 Flash              | gemini-3-flash              | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
+| Grok 4.6                    | grok-4.6                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok 4.5                    | grok-4.5                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok Build 0.1              | grok-build-0.1              | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Qwen3.7 Max                 | qwen3.7-max                 | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                | qwen3.7-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                | qwen3.6-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.5 Plus                | qwen3.5-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| DeepSeek V4 Pro             | deepseek-v4-pro             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash           | deepseek-v4-flash           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M3                  | minimax-m3                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.7                | minimax-m2.7                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.5                | minimax-m2.5                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.2                     | glm-5.2                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.1                     | glm-5.1                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5                       | glm-5                       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.5                   | kimi-k2.5                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                   | kimi-k2.6                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code              | kimi-k2.7-code              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K3                     | kimi-k3                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Big Pickle                  | big-pickle                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5 Free              | mimo-v2.5-free              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Hy3 Free                    | hy3-free                    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Laguna S 2.1 Free           | laguna-s-2.1-free           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3 Ultra Free       | nemotron-3-ultra-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3.5 Lightning Free | nemotron-3.5-lightning-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Free      | deepseek-v4-flash-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |

 [model id](/docs/config/#models) u vašoj OpenCode konfiguraciji koristi format
 `opencode/<model-id>`. Na primjer, za GPT 5.5 u konfiguraciji biste
@@ -147,11 +147,10 @@ Podržavamo pay-as-you-go model. Ispod su cijene **po 1M tokena**.
 | Big Pickle                        | Free   | Free    | Free        | -            |
 | DeepSeek V4 Flash Free            | Free   | Free    | Free        | -            |
 | MiMo-V2.5 Free                    | Free   | Free    | Free        | -            |
+| Hy3 Free                          | Free   | Free    | Free        | -            |
 | Laguna S 2.1 Free                 | Free   | Free    | Free        | -            |
-| Ling-3.0-flash Free               | Free   | Free    | Free        | -            |
-| LongCat-2.0 Free                  | Free   | Free    | Free        | -            |
-| North Mini Code Free              | Free   | Free    | Free        | -            |
 | Nemotron 3 Ultra Free             | Free   | Free    | Free        | -            |
+| Nemotron 3.5 Lightning Free       | Free   | Free    | Free        | -            |
 | MiniMax M3                        | $0.30  | $1.20   | $0.06       | -            |
 | MiniMax M2.7                      | $0.30  | $1.20   | $0.06       | -            |
 | MiniMax M2.5                      | $0.30  | $1.20   | $0.06       | -            |
@@ -185,6 +184,8 @@ Podržavamo pay-as-you-go model. Ispod su cijene **po 1M tokena**.
 | Gemini 3.1 Pro (≤ 200K tokens)    | $2.00  | $12.00  | $0.20       | -            |
 | Gemini 3.1 Pro (> 200K tokens)    | $4.00  | $18.00  | $0.40       | -            |
 | Gemini 3 Flash                    | $0.50  | $3.00   | $0.05       | -            |
+| Grok 4.6 (≤ 200K tokens)          | $2.00  | $6.00   | $0.50       | -            |
+| Grok 4.6 (> 200K tokens)          | $4.00  | $12.00  | $1.00       | -            |
 | Grok 4.5 (≤ 200K tokens)          | $2.00  | $6.00   | $0.30       | -            |
 | Grok 4.5 (> 200K tokens)          | $4.00  | $12.00  | $0.60       | -            |
 | Grok Build 0.1                    | $1.00  | $2.00   | $0.20       | -            |
@@ -224,11 +225,10 @@ Besplatni modeli:

 - DeepSeek V4 Flash Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
 - MiMo-V2.5 Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
+- Hy3 Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
 - Laguna S 2.1 Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
- Ling-3.0-flash Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
- LongCat-2.0 Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
- North Mini Code Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
 - Nemotron 3 Ultra Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
+- Nemotron 3.5 Lightning Free je dostupan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.
 - Big Pickle je stealth model koji je besplatan na OpenCode ograničeno vrijeme. Tim koristi ovo vrijeme da prikupi povratne informacije i poboljša model.

 <a href={email}>Kontaktirajte nas</a> ako imate bilo kakvih pitanja.
@@ -288,11 +288,10 @@ i ne koriste vaše podatke za treniranje modela, uz sljedeće izuzetke:
 - Big Pickle: Tokom besplatnog perioda, prikupljeni podaci mogu se koristiti za poboljšanje modela.
 - DeepSeek V4 Flash Free: Tokom besplatnog perioda, prikupljeni podaci mogu se koristiti za poboljšanje modela.
 - MiMo-V2.5 Free: Tokom besplatnog perioda, prikupljeni podaci mogu se koristiti za poboljšanje modela.
+- Hy3 Free: Tokom besplatnog perioda, prikupljeni podaci mogu se koristiti za poboljšanje modela.
 - Laguna S 2.1 Free: Tokom besplatnog perioda, prikupljeni podaci mogu se koristiti za poboljšanje modela.
- Ling-3.0-flash Free: Tokom besplatnog perioda, prikupljeni podaci mogu se koristiti za poboljšanje modela.
- LongCat-2.0 Free: Tokom besplatnog perioda, prikupljeni podaci mogu se koristiti za poboljšanje modela.
- North Mini Code Free: Tokom besplatnog perioda, prikupljeni podaci mogu biti zadržani i korišteni za poboljšanje modela. Nemojte slati lične ili povjerljive podatke. Pogledajte naše [Uslove korištenja](https://cohere.com/terms-of-use) i [Politiku privatnosti](https://cohere.com/privacy).
 - Nemotron 3 Ultra Free (besplatni NVIDIA endpointi): Samo za probnu upotrebu — nemojte slati lične ili povjerljive podatke. Vaše korištenje se bilježi radi sigurnosti i poboljšanja NVIDIA proizvoda i usluga. Zabilježeni podaci sesije koji se koriste u svrhu poboljšanja nisu povezani s vašim identitetom niti bilo kojim trajnim identifikatorom. Za više informacija o našim praksama obrade podataka pogledajte našu [Politiku privatnosti](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). Interakcijom s ovim endpointom pristajete na naše prikupljanje, bilježenje i korištenje takvih informacija te na [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).
+- Nemotron 3.5 Lightning Free (besplatni NVIDIA endpointi): Samo za probnu upotrebu — nemojte slati lične ili povjerljive podatke. Vaše korištenje se bilježi radi sigurnosti i poboljšanja NVIDIA proizvoda i usluga. Zabilježeni podaci sesije koji se koriste u svrhu poboljšanja nisu povezani s vašim identitetom niti bilo kojim trajnim identifikatorom. Za više informacija o našim praksama obrade podataka pogledajte našu [Politiku privatnosti](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). Interakcijom s ovim endpointom pristajete na naše prikupljanje, bilježenje i korištenje takvih informacija te na [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).
 - OpenAI APIs: Requests are retained for 30 days in accordance with [OpenAI's Data Policies](https://platform.openai.com/docs/guides/your-data).
 - Anthropic APIs: Requests are retained for 30 days in accordance with [Anthropic's Data Policies](https://docs.anthropic.com/en/docs/claude-code/data-usage).

@@ -197,7 +197,7 @@ 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/chat/completions` | `@ai-sdk/openai-compatible` |
+| 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-5.2           | 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` |
@@ -58,69 +58,69 @@ Du bliver opkrævet pr. anmodning, og du kan tilføje kredit til din konto.

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

-| Model                  | Model ID               | Endpoint                                                  | AI SDK-pakke                |
-| ---------------------- | ---------------------- | --------------------------------------------------------- | --------------------------- |
-| GPT 5.6 Sol            | gpt-5.6-sol            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Terra          | gpt-5.6-terra          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Luna           | gpt-5.6-luna           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5                | gpt-5.5                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5 Pro            | gpt-5.5-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4                | gpt-5.4                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Pro            | gpt-5.4-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Mini           | gpt-5.4-mini           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Nano           | gpt-5.4-nano           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex          | gpt-5.3-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex Spark    | gpt-5.3-codex-spark    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2                | gpt-5.2                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2 Codex          | gpt-5.2-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1                | gpt-5.1                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex          | gpt-5.1-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Max      | gpt-5.1-codex-max      | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Mini     | gpt-5.1-codex-mini     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5                  | gpt-5                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Codex            | gpt-5-codex            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Nano             | gpt-5-nano             | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Claude Fable 5         | claude-fable-5         | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 5          | claude-opus-5          | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.8        | claude-opus-4-8        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.7        | claude-opus-4-7        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.6        | claude-opus-4-6        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.5        | claude-opus-4-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 5        | claude-sonnet-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.6      | claude-sonnet-4-6      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.5      | claude-sonnet-4-5      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Haiku 4.5       | claude-haiku-4-5       | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Gemini 3.6 Flash       | gemini-3.6-flash       | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash       | gemini-3.5-flash       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash Lite  | gemini-3.5-flash-lite  | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
-| Gemini 3.1 Pro         | gemini-3.1-pro         | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
-| Gemini 3 Flash         | gemini-3-flash         | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
-| Grok 4.5               | grok-4.5               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Grok Build 0.1         | grok-build-0.1         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Qwen3.7 Max            | qwen3.7-max            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus           | qwen3.7-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus           | qwen3.6-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.5 Plus           | qwen3.5-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| DeepSeek V4 Pro        | deepseek-v4-pro        | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash      | deepseek-v4-flash      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M3             | minimax-m3             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.7           | minimax-m2.7           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.5           | minimax-m2.5           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.2                | glm-5.2                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.1                | glm-5.1                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5                  | glm-5                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.5              | kimi-k2.5              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.6              | kimi-k2.6              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code         | kimi-k2.7-code         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K3                | kimi-k3                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Big Pickle             | big-pickle             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5 Free         | mimo-v2.5-free         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Laguna S 2.1 Free      | laguna-s-2.1-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Ling-3.0-flash Free    | ling-3.0-flash-free    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| LongCat-2.0 Free       | longcat-2.0-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| North Mini Code Free   | north-mini-code-free   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Nemotron 3 Ultra Free  | nemotron-3-ultra-free  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash Free | deepseek-v4-flash-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Model                       | Model ID                    | Endpoint                                                  | AI SDK-pakke                |
+| --------------------------- | --------------------------- | --------------------------------------------------------- | --------------------------- |
+| GPT 5.6 Sol                 | gpt-5.6-sol                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Terra               | gpt-5.6-terra               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                | gpt-5.6-luna                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5                     | gpt-5.5                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5 Pro                 | gpt-5.5-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4                     | gpt-5.4                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Pro                 | gpt-5.4-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Mini                | gpt-5.4-mini                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Nano                | gpt-5.4-nano                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex               | gpt-5.3-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex Spark         | gpt-5.3-codex-spark         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2                     | gpt-5.2                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2 Codex               | gpt-5.2-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1                     | gpt-5.1                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex               | gpt-5.1-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Max           | gpt-5.1-codex-max           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Mini          | gpt-5.1-codex-mini          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5                       | gpt-5                       | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Codex                 | gpt-5-codex                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Nano                  | gpt-5-nano                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Claude Fable 5              | claude-fable-5              | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 5               | claude-opus-5               | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.8             | claude-opus-4-8             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.7             | claude-opus-4-7             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.6             | claude-opus-4-6             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.5             | claude-opus-4-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 5             | claude-sonnet-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.6           | claude-sonnet-4-6           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.5           | claude-sonnet-4-5           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Haiku 4.5            | claude-haiku-4-5            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Gemini 3.6 Flash            | gemini-3.6-flash            | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash            | gemini-3.5-flash            | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash Lite       | gemini-3.5-flash-lite       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
+| Gemini 3.1 Pro              | gemini-3.1-pro              | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
+| Gemini 3 Flash              | gemini-3-flash              | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
+| Grok 4.6                    | grok-4.6                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok 4.5                    | grok-4.5                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok Build 0.1              | grok-build-0.1              | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Qwen3.7 Max                 | qwen3.7-max                 | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                | qwen3.7-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                | qwen3.6-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.5 Plus                | qwen3.5-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| DeepSeek V4 Pro             | deepseek-v4-pro             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash           | deepseek-v4-flash           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M3                  | minimax-m3                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.7                | minimax-m2.7                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.5                | minimax-m2.5                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.2                     | glm-5.2                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.1                     | glm-5.1                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5                       | glm-5                       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.5                   | kimi-k2.5                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                   | kimi-k2.6                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code              | kimi-k2.7-code              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K3                     | kimi-k3                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Big Pickle                  | big-pickle                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5 Free              | mimo-v2.5-free              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Hy3 Free                    | hy3-free                    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Laguna S 2.1 Free           | laguna-s-2.1-free           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3 Ultra Free       | nemotron-3-ultra-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3.5 Lightning Free | nemotron-3.5-lightning-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Free      | deepseek-v4-flash-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |

 [model id](/docs/config/#models) i din OpenCode-konfiguration
 bruger formatet `opencode/<model-id>`. For eksempel ville du for GPT 5.5
@@ -147,11 +147,10 @@ Vi understøtter en pay-as-you-go-model. Nedenfor er priserne **pr. 1M tokens**.
 | Big Pickle                        | Free   | Free    | Free        | -            |
 | DeepSeek V4 Flash Free            | Free   | Free    | Free        | -            |
 | MiMo-V2.5 Free                    | Free   | Free    | Free        | -            |
+| Hy3 Free                          | Free   | Free    | Free        | -            |
 | Laguna S 2.1 Free                 | Free   | Free    | Free        | -            |
-| Ling-3.0-flash Free               | Free   | Free    | Free        | -            |
-| LongCat-2.0 Free                  | Free   | Free    | Free        | -            |
-| North Mini Code Free              | Free   | Free    | Free        | -            |
 | Nemotron 3 Ultra Free             | Free   | Free    | Free        | -            |
+| Nemotron 3.5 Lightning Free       | Free   | Free    | Free        | -            |
 | MiniMax M3                        | $0.30  | $1.20   | $0.06       | -            |
 | MiniMax M2.7                      | $0.30  | $1.20   | $0.06       | -            |
 | MiniMax M2.5                      | $0.30  | $1.20   | $0.06       | -            |
@@ -185,6 +184,8 @@ Vi understøtter en pay-as-you-go-model. Nedenfor er priserne **pr. 1M tokens**.
 | Gemini 3.1 Pro (≤ 200K tokens)    | $2.00  | $12.00  | $0.20       | -            |
 | Gemini 3.1 Pro (> 200K tokens)    | $4.00  | $18.00  | $0.40       | -            |
 | Gemini 3 Flash                    | $0.50  | $3.00   | $0.05       | -            |
+| Grok 4.6 (≤ 200K tokens)          | $2.00  | $6.00   | $0.50       | -            |
+| Grok 4.6 (> 200K tokens)          | $4.00  | $12.00  | $1.00       | -            |
 | Grok 4.5 (≤ 200K tokens)          | $2.00  | $6.00   | $0.30       | -            |
 | Grok 4.5 (> 200K tokens)          | $4.00  | $12.00  | $0.60       | -            |
 | Grok Build 0.1                    | $1.00  | $2.00   | $0.20       | -            |
@@ -224,11 +225,10 @@ De gratis modeller:

 - DeepSeek V4 Flash Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
 - MiMo-V2.5 Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
+- Hy3 Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
 - Laguna S 2.1 Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
- Ling-3.0-flash Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
- LongCat-2.0 Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
- North Mini Code Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
 - Nemotron 3 Ultra Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
+- Nemotron 3.5 Lightning Free er tilgængelig på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.
 - Big Pickle er en stealth-model, som er gratis på OpenCode i en begrænset periode. Teamet bruger denne tid til at indsamle feedback og forbedre modellen.

 <a href={email}>Kontakt os</a>, hvis du har spørgsmål.
@@ -286,11 +286,10 @@ Alle vores modeller hostes i US. Vores udbydere følger en nul-opbevaringspoliti
 - Big Pickle: I den gratis periode kan indsamlede data blive brugt til at forbedre modellen.
 - DeepSeek V4 Flash Free: I den gratis periode kan indsamlede data blive brugt til at forbedre modellen.
 - MiMo-V2.5 Free: I den gratis periode kan indsamlede data blive brugt til at forbedre modellen.
+- Hy3 Free: I den gratis periode kan indsamlede data blive brugt til at forbedre modellen.
 - Laguna S 2.1 Free: I den gratis periode kan indsamlede data blive brugt til at forbedre modellen.
- Ling-3.0-flash Free: I den gratis periode kan indsamlede data blive brugt til at forbedre modellen.
- LongCat-2.0 Free: I den gratis periode kan indsamlede data blive brugt til at forbedre modellen.
- North Mini Code Free: I gratisperioden kan indsamlede data blive opbevaret og brugt til at forbedre modellen. Indsend ikke personlige eller fortrolige oplysninger. Se vores [Brugsvilkår](https://cohere.com/terms-of-use) og [Privatlivspolitik](https://cohere.com/privacy).
 - Nemotron 3 Ultra Free (gratis NVIDIA-endpoints): Kun til prøvebrug — indsend ikke personlige eller fortrolige data. Din brug logges af sikkerhedshensyn og for at forbedre NVIDIAs produkter og tjenester. De loggede sessionsdata, der bruges til forbedringsformål, er ikke knyttet til din identitet eller nogen vedvarende identifikator. For mere information om vores databehandlingspraksis, se vores [privatlivspolitik](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). Ved at interagere med dette endpoint giver du samtykke til vores indsamling, registrering og brug af sådanne oplysninger samt [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).
+- Nemotron 3.5 Lightning Free (gratis NVIDIA-endpoints): Kun til prøvebrug — indsend ikke personlige eller fortrolige data. Din brug logges af sikkerhedshensyn og for at forbedre NVIDIAs produkter og tjenester. De loggede sessionsdata, der bruges til forbedringsformål, er ikke knyttet til din identitet eller nogen vedvarende identifikator. For mere information om vores databehandlingspraksis, se vores [privatlivspolitik](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). Ved at interagere med dette endpoint giver du samtykke til vores indsamling, registrering og brug af sådanne oplysninger samt [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).
 - OpenAI APIs: Anmodninger opbevares i 30 dage i overensstemmelse med [OpenAI's Data Policies](https://platform.openai.com/docs/guides/your-data).
 - Anthropic APIs: Anmodninger opbevares i 30 dage i overensstemmelse med [Anthropic's Data Policies](https://docs.anthropic.com/en/docs/claude-code/data-usage).

@@ -187,7 +187,7 @@ 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/chat/completions` | `@ai-sdk/openai-compatible` |
+| 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-5.2           | 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` |
@@ -49,69 +49,69 @@ Dir wird pro Anfrage berechnet, und du kannst deinem Konto Guthaben hinzufügen.

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

-| Model                  | Model ID               | Endpoint                                                  | AI SDK Package              |
-| ---------------------- | ---------------------- | --------------------------------------------------------- | --------------------------- |
-| GPT 5.6 Sol            | gpt-5.6-sol            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Terra          | gpt-5.6-terra          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.6 Luna           | gpt-5.6-luna           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5                | gpt-5.5                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.5 Pro            | gpt-5.5-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4                | gpt-5.4                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Pro            | gpt-5.4-pro            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Mini           | gpt-5.4-mini           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.4 Nano           | gpt-5.4-nano           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex          | gpt-5.3-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.3 Codex Spark    | gpt-5.3-codex-spark    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2                | gpt-5.2                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.2 Codex          | gpt-5.2-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1                | gpt-5.1                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex          | gpt-5.1-codex          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Max      | gpt-5.1-codex-max      | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5.1 Codex Mini     | gpt-5.1-codex-mini     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5                  | gpt-5                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Codex            | gpt-5-codex            | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| GPT 5 Nano             | gpt-5-nano             | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Claude Fable 5         | claude-fable-5         | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 5          | claude-opus-5          | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.8        | claude-opus-4-8        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.7        | claude-opus-4-7        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.6        | claude-opus-4-6        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Opus 4.5        | claude-opus-4-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 5        | claude-sonnet-5        | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.6      | claude-sonnet-4-6      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Sonnet 4.5      | claude-sonnet-4-5      | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Claude Haiku 4.5       | claude-haiku-4-5       | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Gemini 3.6 Flash       | gemini-3.6-flash       | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash       | gemini-3.5-flash       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
-| Gemini 3.5 Flash Lite  | gemini-3.5-flash-lite  | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
-| Gemini 3.1 Pro         | gemini-3.1-pro         | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
-| Gemini 3 Flash         | gemini-3-flash         | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
-| Grok 4.5               | grok-4.5               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Grok Build 0.1         | grok-build-0.1         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
-| Qwen3.7 Max            | qwen3.7-max            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.7 Plus           | qwen3.7-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.6 Plus           | qwen3.6-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| Qwen3.5 Plus           | qwen3.5-plus           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
-| DeepSeek V4 Pro        | deepseek-v4-pro        | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash      | deepseek-v4-flash      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M3             | minimax-m3             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.7           | minimax-m2.7           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiniMax M2.5           | minimax-m2.5           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.2                | glm-5.2                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5.1                | glm-5.1                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| GLM 5                  | glm-5                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.5              | kimi-k2.5              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.6              | kimi-k2.6              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K2.7 Code         | kimi-k2.7-code         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Kimi K3                | kimi-k3                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Big Pickle             | big-pickle             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5 Free         | mimo-v2.5-free         | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Laguna S 2.1 Free      | laguna-s-2.1-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Ling-3.0-flash Free    | ling-3.0-flash-free    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| LongCat-2.0 Free       | longcat-2.0-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| North Mini Code Free   | north-mini-code-free   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| Nemotron 3 Ultra Free  | nemotron-3-ultra-free  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
-| DeepSeek V4 Flash Free | deepseek-v4-flash-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Model                       | Model ID                    | Endpoint                                                  | AI SDK Package              |
+| --------------------------- | --------------------------- | --------------------------------------------------------- | --------------------------- |
+| GPT 5.6 Sol                 | gpt-5.6-sol                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Terra               | gpt-5.6-terra               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.6 Luna                | gpt-5.6-luna                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5                     | gpt-5.5                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.5 Pro                 | gpt-5.5-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4                     | gpt-5.4                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Pro                 | gpt-5.4-pro                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Mini                | gpt-5.4-mini                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.4 Nano                | gpt-5.4-nano                | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex               | gpt-5.3-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.3 Codex Spark         | gpt-5.3-codex-spark         | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2                     | gpt-5.2                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.2 Codex               | gpt-5.2-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1                     | gpt-5.1                     | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex               | gpt-5.1-codex               | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Max           | gpt-5.1-codex-max           | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5.1 Codex Mini          | gpt-5.1-codex-mini          | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5                       | gpt-5                       | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Codex                 | gpt-5-codex                 | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| GPT 5 Nano                  | gpt-5-nano                  | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Claude Fable 5              | claude-fable-5              | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 5               | claude-opus-5               | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.8             | claude-opus-4-8             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.7             | claude-opus-4-7             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.6             | claude-opus-4-6             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Opus 4.5             | claude-opus-4-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 5             | claude-sonnet-5             | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.6           | claude-sonnet-4-6           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Sonnet 4.5           | claude-sonnet-4-5           | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Claude Haiku 4.5            | claude-haiku-4-5            | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Gemini 3.6 Flash            | gemini-3.6-flash            | `https://opencode.ai/zen/v1/models/gemini-3.6-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash            | gemini-3.5-flash            | `https://opencode.ai/zen/v1/models/gemini-3.5-flash`      | `@ai-sdk/google`            |
+| Gemini 3.5 Flash Lite       | gemini-3.5-flash-lite       | `https://opencode.ai/zen/v1/models/gemini-3.5-flash-lite` | `@ai-sdk/google`            |
+| Gemini 3.1 Pro              | gemini-3.1-pro              | `https://opencode.ai/zen/v1/models/gemini-3.1-pro`        | `@ai-sdk/google`            |
+| Gemini 3 Flash              | gemini-3-flash              | `https://opencode.ai/zen/v1/models/gemini-3-flash`        | `@ai-sdk/google`            |
+| Grok 4.6                    | grok-4.6                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok 4.5                    | grok-4.5                    | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Grok Build 0.1              | grok-build-0.1              | `https://opencode.ai/zen/v1/responses`                    | `@ai-sdk/openai`            |
+| Qwen3.7 Max                 | qwen3.7-max                 | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.7 Plus                | qwen3.7-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.6 Plus                | qwen3.6-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| Qwen3.5 Plus                | qwen3.5-plus                | `https://opencode.ai/zen/v1/messages`                     | `@ai-sdk/anthropic`         |
+| DeepSeek V4 Pro             | deepseek-v4-pro             | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash           | deepseek-v4-flash           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M3                  | minimax-m3                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.7                | minimax-m2.7                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiniMax M2.5                | minimax-m2.5                | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.2                     | glm-5.2                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5.1                     | glm-5.1                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| GLM 5                       | glm-5                       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.5                   | kimi-k2.5                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.6                   | kimi-k2.6                   | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K2.7 Code              | kimi-k2.7-code              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Kimi K3                     | kimi-k3                     | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Big Pickle                  | big-pickle                  | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5 Free              | mimo-v2.5-free              | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Hy3 Free                    | hy3-free                    | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Laguna S 2.1 Free           | laguna-s-2.1-free           | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3 Ultra Free       | nemotron-3-ultra-free       | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| Nemotron 3.5 Lightning Free | nemotron-3.5-lightning-free | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |
+| DeepSeek V4 Flash Free      | deepseek-v4-flash-free      | `https://opencode.ai/zen/v1/chat/completions`             | `@ai-sdk/openai-compatible` |

 Die [Model-ID](/docs/config/#models) in deiner OpenCode-Konfiguration verwendet das Format `opencode/<model-id>`. Für GPT 5.5 würdest du zum Beispiel `opencode/gpt-5.5` in deiner Konfiguration verwenden.

@@ -136,11 +136,10 @@ Wir unterstützen ein Pay-as-you-go-Modell. Unten findest du die Preise **pro 1M
 | Big Pickle                        | Free   | Free    | Free        | -            |
 | DeepSeek V4 Flash Free            | Free   | Free    | Free        | -            |
 | MiMo-V2.5 Free                    | Free   | Free    | Free        | -            |
+| Hy3 Free                          | Free   | Free    | Free        | -            |
 | Laguna S 2.1 Free                 | Free   | Free    | Free        | -            |
-| Ling-3.0-flash Free               | Free   | Free    | Free        | -            |
-| LongCat-2.0 Free                  | Free   | Free    | Free        | -            |
-| North Mini Code Free              | Free   | Free    | Free        | -            |
 | Nemotron 3 Ultra Free             | Free   | Free    | Free        | -            |
+| Nemotron 3.5 Lightning Free       | Free   | Free    | Free        | -            |
 | MiniMax M3                        | $0.30  | $1.20   | $0.06       | -            |
 | MiniMax M2.7                      | $0.30  | $1.20   | $0.06       | -            |
 | MiniMax M2.5                      | $0.30  | $1.20   | $0.06       | -            |
@@ -174,6 +173,8 @@ Wir unterstützen ein Pay-as-you-go-Modell. Unten findest du die Preise **pro 1M
 | Gemini 3.1 Pro (≤ 200K tokens)    | $2.00  | $12.00  | $0.20       | -            |
 | Gemini 3.1 Pro (> 200K tokens)    | $4.00  | $18.00  | $0.40       | -            |
 | Gemini 3 Flash                    | $0.50  | $3.00   | $0.05       | -            |
+| Grok 4.6 (≤ 200K tokens)          | $2.00  | $6.00   | $0.50       | -            |
+| Grok 4.6 (> 200K tokens)          | $4.00  | $12.00  | $1.00       | -            |
 | Grok 4.5 (≤ 200K tokens)          | $2.00  | $6.00   | $0.30       | -            |
 | Grok 4.5 (> 200K tokens)          | $4.00  | $12.00  | $0.60       | -            |
 | Grok Build 0.1                    | $1.00  | $2.00   | $0.20       | -            |
@@ -213,11 +214,10 @@ Die kostenlosen Modelle:

 - DeepSeek V4 Flash Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
 - MiMo-V2.5 Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
+- Hy3 Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
 - Laguna S 2.1 Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
- Ling-3.0-flash Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
- LongCat-2.0 Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
- North Mini Code Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
 - Nemotron 3 Ultra Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
+- Nemotron 3.5 Lightning Free ist für begrenzte Zeit auf OpenCode verfügbar. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.
 - Big Pickle ist ein Stealth-Modell, das für begrenzte Zeit kostenlos auf OpenCode verfügbar ist. Das Team nutzt diese Zeit, um Feedback zu sammeln und das Modell zu verbessern.

 <a href={email}>Kontaktiere uns</a>, wenn du Fragen hast.
@@ -272,11 +272,10 @@ Alle unsere Modelle werden in den USA gehostet. Unsere Provider folgen einer Zer
 - Big Pickle: Während des kostenlosen Zeitraums können gesammelte Daten zur Verbesserung des Modells verwendet werden.
 - DeepSeek V4 Flash Free: Während des kostenlosen Zeitraums können gesammelte Daten zur Verbesserung des Modells verwendet werden.
 - MiMo-V2.5 Free: Während des kostenlosen Zeitraums können gesammelte Daten zur Verbesserung des Modells verwendet werden.
+- Hy3 Free: Während des kostenlosen Zeitraums können gesammelte Daten zur Verbesserung des Modells verwendet werden.
 - Laguna S 2.1 Free: Während des kostenlosen Zeitraums können gesammelte Daten zur Verbesserung des Modells verwendet werden.
- Ling-3.0-flash Free: Während des kostenlosen Zeitraums können gesammelte Daten zur Verbesserung des Modells verwendet werden.
- LongCat-2.0 Free: Während des kostenlosen Zeitraums können gesammelte Daten zur Verbesserung des Modells verwendet werden.
- North Mini Code Free: Während des kostenlosen Zeitraums können erhobene Daten gespeichert und zur Verbesserung des Modells verwendet werden. Übermitteln Sie keine personenbezogenen oder vertraulichen Daten. Weitere Informationen finden Sie in unseren [Nutzungsbedingungen](https://cohere.com/terms-of-use) und unserer [Datenschutzerklärung](https://cohere.com/privacy).
 - Nemotron 3 Ultra Free (kostenlose NVIDIA-Endpunkte): Nur für Testzwecke — übermitteln Sie keine personenbezogenen oder vertraulichen Daten. Ihre Nutzung wird zu Sicherheitszwecken und zur Verbesserung der Produkte und Dienste von NVIDIA protokolliert. Die zu Verbesserungszwecken protokollierten Sitzungsdaten sind nicht mit Ihrer Identität oder einem dauerhaften Identifikator verknüpft. Weitere Informationen zu unseren Datenverarbeitungspraktiken finden Sie in unserer [Datenschutzrichtlinie](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). Durch die Interaktion mit diesem Endpunkt stimmen Sie unserer Erhebung, Aufzeichnung und Nutzung solcher Informationen sowie den [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf) zu.
+- Nemotron 3.5 Lightning Free (kostenlose NVIDIA-Endpunkte): Nur für Testzwecke — übermitteln Sie keine personenbezogenen oder vertraulichen Daten. Ihre Nutzung wird zu Sicherheitszwecken und zur Verbesserung der Produkte und Dienste von NVIDIA protokolliert. Die zu Verbesserungszwecken protokollierten Sitzungsdaten sind nicht mit Ihrer Identität oder einem dauerhaften Identifikator verknüpft. Weitere Informationen zu unseren Datenverarbeitungspraktiken finden Sie in unserer [Datenschutzrichtlinie](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). Durch die Interaktion mit diesem Endpunkt stimmen Sie unserer Erhebung, Aufzeichnung und Nutzung solcher Informationen sowie den [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf) zu.
 - OpenAI APIs: Anfragen werden in Übereinstimmung mit [OpenAI's Data Policies](https://platform.openai.com/docs/guides/your-data) 30 Tage lang gespeichert.
 - Anthropic APIs: Anfragen werden in Übereinstimmung mit [Anthropic's Data Policies](https://docs.anthropic.com/en/docs/claude-code/data-usage) 30 Tage lang gespeichert.

--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Aiden Cline	9e7b5f1701	feat(opencode): cache-friendly compaction via primary loop request path Default compaction no longer uses a dedicated compaction agent. It inherits the session's active agent and delegates request assembly to the primary loop's shared envelope (same system prompt, instructions, skills, MCP context, and tool definitions), replaying the typed message head and appending the summary instruction as the final user message. This keeps the compaction request prefix-identical to the preceding turn so provider prompt/KV caches are reused instead of fully re-prefilled. An explicitly configured agent.compaction preserves the legacy behavior exactly: dedicated hidden agent, optional model override, and the serialized-transcript summary request.	2026-08-13 23:10:07 -05:00
Jack	d0c2b41adf	docs(go): use responses API for Grok 4.5 (#42373 )	2026-08-14 01:28:55 +08:00
Frank	c7af47f9ed	update grok endpoint	2026-08-13 13:27:41 -04:00
Kit Langton	6c035e1fd7	fix(core): preserve unicode in grep previews (#42356 )	2026-08-13 16:51:43 +00:00
opencode-agent[bot]	ab7cbc808f	chore: generate	2026-08-13 16:30:12 +00:00
Aditya Sethi	62387f39d4	fix(skills): Update global config path in documentation (#42337 )	2026-08-13 18:27:29 +02:00
Jack	864889ab9f	docs: remove Ling 3.0 Tiny free model (#42314 )	2026-08-13 20:48:54 +08:00
opencode	cc4b456129	sync release versions for v1.18.18	2026-08-13 01:15:01 +00:00
opencode-agent[bot]	14b37df391	chore: update nix node_modules hashes	2026-08-12 22:36:40 +00:00
opencode-agent[bot]	91df883231	fix(opencode): select Kimi prompt by provider (#42161 ) Co-authored-by: Aiden Cline <rekram1-node@users.noreply.github.com>	2026-08-12 17:29:03 -05:00
opencode-agent[bot]	6fea419feb	fix(groq): pass through reasoning effort (#42166 ) Co-authored-by: Aiden Cline <rekram1-node@users.noreply.github.com>	2026-08-12 17:20:23 -05:00
opencode-agent[bot]	beeabe2e4b	fix(mistral): pass through reasoning effort (#42164 ) Co-authored-by: Aiden Cline <rekram1-node@users.noreply.github.com>	2026-08-12 17:20:07 -05:00
opencode-agent[bot]	502310f4df	fix(xai): pass through reasoning effort (#42160 ) Co-authored-by: Aiden Cline <rekram1-node@users.noreply.github.com>	2026-08-12 17:19:17 -05:00
opencode	37fe5c83dc	sync release versions for v1.18.17	2026-08-12 20:25:04 +00:00
Aiden Cline	dab2637217	fix(compaction): adjust instructions and structure to be more clear to smaller models like dsv4 flash (#42045 ) Co-authored-by: akenra <37288280+akenra@users.noreply.github.com>	2026-08-12 14:36:51 -05:00
opencode-agent[bot]	39fb919a05	chore: add neriousy to team members (#42107 ) Co-authored-by: Aiden Cline <rekram1-node@users.noreply.github.com>	2026-08-12 13:30:38 -05:00
opencode-agent[bot]	999be62662	chore: generate	2026-08-12 16:56:04 +00:00
opencode-agent[bot]	521906f5fa	docs(go): clarify DeepSeek ZDR coverage (#42085 ) Co-authored-by: Dax Raad <mail@thdxr.com>	2026-08-12 12:53:58 -04:00
Frank	069165eb5b	Merge branch 'dev' of github.com:anomalyco/opencode into dev	2026-08-12 12:48:10 -04:00
Frank	df09c3ec61	update ds v4 pro	2026-08-12 12:48:08 -04:00
opencode-agent[bot]	6d3ae4d63d	chore: generate	2026-08-12 16:47:08 +00:00
Adam	284187ac55	fix(ci): authenticate pulumi downloads	2026-08-12 11:45:00 -05:00
Adam	7e0353cca9	fix(stats): correct r2 daily totals	2026-08-12 11:36:23 -05:00
SKY ZHAO	959c8bd498	docs: fix provider display name and PAT typos (#42034 ) Co-authored-by: skyzhao1223 <skyzhao1223@users.noreply.github.com>	2026-08-12 10:32:23 -05:00
SKY ZHAO	ca3df21b7f	docs: fix broken DigitalOcean and Daytona links (#42048 ) Co-authored-by: skyzhao1223 <skyzhao1223@users.noreply.github.com>	2026-08-12 10:31:47 -05:00
Matthew Feroz	8571a922db	fix(provider): add Merge Gateway reasoning variants (#41867 )	2026-08-12 10:31:19 -05:00
Frank	d92d1e654b	docs(zen): add Grok 4.6	2026-08-12 10:53:38 -04:00
Adam	46a14e685a	feat(stats): query r2 data catalog	2026-08-12 09:49:10 -05:00
Jack	1f94d8a3c8	docs(zen): remove expired free models (#41943 )	2026-08-12 12:03:45 +08:00
Aiden Cline	c78986831c	fix(opencode): cap session retries with jitter (#41939 )	2026-08-11 22:50:52 -05:00
vimtor	d470434746	refactor(console): simplify go usage response	2026-08-11 23:18:57 +02:00
Steven Ao	561afb401a	fix(opencode): detect Copilot PDF input support (#41522 )	2026-08-11 16:04:36 -05:00
Arif Rahman (Bolt)	2b8a5969e9	feat(console): add go usage endpoint (#16513 ) Co-authored-by: vimtor <vn4varro@gmail.com>	2026-08-11 22:25:23 +02:00
Jack	36b205370d	docs(zen): add Hy3 Free (#41814 )	2026-08-12 02:42:06 +08:00
opencode-agent[bot]	6afef2f881	fix(console): update blocked account message (#41819 ) Co-authored-by: Dax Raad <d@ironbay.co> Co-authored-by: Dax <mail@thdxr.com>	2026-08-11 14:41:37 -04:00
opencode-agent[bot]	0d927ba03f	chore: generate	2026-08-11 14:17:42 +00:00
Jack	9fdd4824d3	docs(zen): add Nemotron 3.5 Lightning (#41750 )	2026-08-11 22:15:53 +08:00
opencode-agent[bot]	d041eee55c	chore: use DeepSeek for beta conflicts (#41669 ) Co-authored-by: Luke Parker <10430890+Hona@users.noreply.github.com>	2026-08-11 04:58:15 +00:00
Luke Parker	d62e50075a	chore: build beta from v2 (#41627 )	2026-08-11 14:33:42 +10:00
opencode-agent[bot]	5d953482ab	fix(provider): scope DeepSeek V4 Flash sampling defaults (#41620 ) Co-authored-by: Aiden Cline <63023139+rekram1-node@users.noreply.github.com> Co-authored-by: 侯秀凯 <9627034+hoooou@users.noreply.github.com>	2026-08-10 23:01:56 -05:00
OpeOginni	3a90639cb5	fix(ui): correct OC-2 weak icon color (#41504 )	2026-08-10 17:08:45 -05:00
Matthias Reso	b9f3b382fc	fix(session): route all Muse family models to the Meta system prompt (#41581 )	2026-08-10 13:28:04 -05:00
opencode-agent[bot]	550d1ffd24	fix(i18n): use widely recognized developer terminology (#41532 ) Co-authored-by: Luke Parker <10430890+Hona@users.noreply.github.com>	2026-08-10 22:00:00 +10:00
Brendan Allan	941e71dbbb	fix(app): use current default model (#38603 )	2026-08-10 15:14:17 +08:00
opencode	d90532a595	sync release versions for v1.18.16	2026-08-10 06:07:05 +00:00
opencode-agent[bot]	0bff28de09	fix(stats): fall back after full sync failure (#41411 )	2026-08-09 10:58:58 -05:00
opencode-agent[bot]	38e10eb140	fix(opencode): ignore unknown config fields (#41312 )	2026-08-08 14:53:43 -04:00
opencode-agent[bot]	fe82a1b6ca	chore: generate	2026-08-08 03:39:08 +00:00
Brendan Allan	2ea728f731	fix(app): populate project picker from home (#41158 )	2026-08-08 11:37:30 +08:00
Aarav Sareen	284214c78d	feat(app): open project menu on right-click (#41013 )	2026-08-07 17:25:46 +08:00
日常无语的无语者	b05a0d1fbe	fix(i18n): use 词元 instead of 令牌 for token in zh locale (#40977 )	2026-08-07 19:22:23 +10:00
Jack	be25c905fa	docs(zen): replace Ling Flash with Tiny (#41027 )	2026-08-07 16:48:42 +08:00
Brendan Allan	5347b5e009	fix(app): register new home projects (#41018 )	2026-08-07 07:34:32 +00:00
opencode-agent[bot]	b467518dae	fix(desktop): preserve macOS app on window close (#40974 ) Co-authored-by: Brendan Allan <14191578+Brendonovich@users.noreply.github.com>	2026-08-07 07:25:12 +00:00
opencode	1ec6bdc8c6	sync release versions for v1.18.15	2026-08-07 06:49:53 +00:00
Aarav Sareen	265472ef72	fix(ui): use tokens for button loading state (#40884 )	2026-08-07 12:11:00 +05:30