feat(opencode): route cloudflare ai gateway openai and anthropic models through native passthroughs

Co-authored-by: Keefe Tang <keefe@pentagram.me>
Merge branch 'dev' into fix/ai-gateway-scope-token
2026-08-14 15:32:52 -04:00 · 2026-08-14 13:31:13 -05:00 · 2026-08-13 22:14:38 -05:00 · 2026-08-11 23:59:55 -05:00 · 2026-06-24 11:52:13 +10:00
47 changed files with 290 additions and 264 deletions
@@ -1,8 +1,8 @@
 {
  "nodeModules": {
-    "x86_64-linux": "sha256-kDCnJMnaK/Jq7ckcpPB7Vl9v98EMSdcehZAtf8jNjTs=",
-    "aarch64-linux": "sha256-0aR+OJGXS5HMlXbe/BHybjIRvdNJJw6gjW+jr6Dk7Pk=",
-    "aarch64-darwin": "sha256-loLrV6xiorhwS/N2hlpiKSKX172Qxy+auNiPzFBhQSc=",
-    "x86_64-darwin": "sha256-PNEpQBLAz8M274bSyTpp0jofETn2L+D0uBiJHUV7nB0="
+    "x86_64-linux": "sha256-TNwKfqxD83UpZuCKN8FdEWN+CcQUP9CkCQSLGNqR/sA=",
+    "aarch64-linux": "sha256-qzvOJZzmq2QhlauElw8GwgQnCPHdhexI52L0md5zrxQ=",
+    "aarch64-darwin": "sha256-ZzoyLayOFfcYUAg35ZbZ2WapxDdd9IUWqy2xkxZH4QM=",
+    "x86_64-darwin": "sha256-maP/qLeaC3q8VcmNIPyIKlnplxFXJ7ULho3v21/16Mw="
  }
 }
@@ -25,7 +25,6 @@ const checkLoggedIn = query(async () => {
 const models = [
  { name: "Grok 4.5", training: "go.faq.a5.notUsed", retention: "go.faq.a5.retention30" },
  { name: "GPT 5.6 Luna", training: "go.faq.a5.notUsed", retention: "go.faq.a5.retention30" },
-  { name: "GLM-5.3", training: "go.faq.a5.notUsed", retention: "go.faq.a5.retention0" },
  { name: "GLM-5.2", training: "go.faq.a5.notUsed", retention: "go.faq.a5.retention0" },
  { name: "GLM-5.1", training: "go.faq.a5.notUsed", retention: "go.faq.a5.retention0" },
  { name: "Kimi K3", training: "go.faq.a5.notUsed", retention: "go.faq.a5.retention0" },
@@ -306,7 +306,6 @@ export function LiteSection(props: { lite: LiteSubscription | undefined }) {
          <ul data-slot="promo-models">
            <li>Grok 4.5</li>
            <li>GPT 5.6 Luna</li>
-            <li>GLM-5.3</li>
            <li>GLM-5.2</li>
            <li>GLM-5.1</li>
            <li>Kimi K3</li>
@@ -24,9 +24,15 @@ export const CloudflareAIGatewayPlugin = define({
          apiKey: config.apiKey,
          options: gatewayOptions(evt.options, metadata),
        } as any)
-        const unified = createUnified({ apiKey: config.apiKey })
        evt.sdk = {
          languageModel(modelID: string) {
+            // Workers AI is the only first-party provider whose upstream is Cloudflare itself, so it is
+            // the only one that should receive the Cloudflare token as its upstream Authorization header.
+            // The Unified API addresses Workers AI both with the explicit "workers-ai/" prefix and as
+            // bare "@cf/..." ids. Third-party providers must not receive the token; they rely on the
+            // gateway's stored/BYOK keys instead.
+            const isWorkersAi = modelID.startsWith("workers-ai/") || modelID.startsWith("@cf/")
+            const unified = createUnified(isWorkersAi ? { apiKey: config.apiKey } : {})
            return gateway(unified(modelID))
          },
        }
@@ -61,16 +61,5 @@ export async function CloudflareAIGatewayAuthPlugin(_input: PluginInput): Promis
        },
      ],
    },
-    "chat.params": async (input, output) => {
-      if (input.model.providerID !== "cloudflare-ai-gateway") return
-      // The unified gateway routes through @ai-sdk/openai-compatible, which
-      // always emits max_tokens. OpenAI reasoning models (gpt-5.x, o-series)
-      // reject that field and require max_completion_tokens instead, and the
-      // compatible SDK has no way to rename it. Drop the cap so OpenAI falls
-      // back to the model's default output budget.
-      if (!input.model.api.id.toLowerCase().startsWith("openai/")) return
-      if (!input.model.capabilities.reasoning) return
-      output.maxOutputTokens = undefined
-    },
  }
 }
@@ -800,9 +800,10 @@ function custom(dep: CustomDep): Record<string, CustomLoader> {
        )
      }

-      // Use official ai-gateway-provider package (v2.x for AI SDK v5 compatibility)
      const { createAiGateway } = yield* Effect.promise(() => import("ai-gateway-provider"))
      const { createUnified } = yield* Effect.promise(() => import("ai-gateway-provider/providers/unified"))
+      const { createOpenAI } = yield* Effect.promise(() => import("ai-gateway-provider/providers/openai"))
+      const { createAnthropic } = yield* Effect.promise(() => import("ai-gateway-provider/providers/anthropic"))

      const metadata = iife(() => {
        if (input.options?.metadata) return input.options.metadata
@@ -829,12 +830,24 @@ function custom(dep: CustomDep): Record<string, CustomLoader> {
        apiKey: apiToken,
        ...(Object.values(opts).some((v) => v !== undefined) ? { options: opts } : {}),
      })
-      const unified = createUnified({ apiKey: apiToken })
-
      return {
        autoload: true,
        async getModel(_sdk: any, modelID: string, _options?: Record<string, any>) {
-          // Model IDs use Unified API format: provider/model (e.g., "anthropic/claude-sonnet-4-5")
+          // Model IDs use Unified API format: provider/model (e.g., "anthropic/claude-sonnet-4-5").
+          // OpenAI and Anthropic ride their native passthrough routes so agents get the Responses
+          // and Messages APIs; new OpenAI models reject tools+reasoning_effort on chat completions.
+          // The passthrough wrappers inject a CF_TEMP_TOKEN sentinel that the gateway strips before
+          // dispatch, so upstream billing stays on the gateway (Unified Billing / stored BYOK).
+          if (modelID.startsWith("openai/")) return aigateway(createOpenAI()(modelID.slice("openai/".length)))
+          if (modelID.startsWith("anthropic/"))
+            return aigateway(createAnthropic()(modelID.slice("anthropic/".length)))
+          // Workers AI is the only first-party provider whose upstream is Cloudflare itself, so it is
+          // the only one that should receive the Cloudflare token as its upstream Authorization header.
+          // The Unified API addresses Workers AI both with the explicit "workers-ai/" prefix and as
+          // bare "@cf/..." ids. Third-party providers must not receive the token; they rely on the
+          // gateway's stored/BYOK keys instead.
+          const isWorkersAi = modelID.startsWith("workers-ai/") || modelID.startsWith("@cf/")
+          const unified = createUnified(isWorkersAi ? { apiKey: apiToken } : {})
          return aigateway(unified(modelID))
        },
        options: {},
@@ -1209,6 +1222,17 @@ function cost(c: ModelsDev.Model["cost"]): Model["cost"] {
  return result
 }

+// Cloudflare AI Gateway routes OpenAI and Anthropic models through their native
+// passthrough SDKs (Responses / Messages APIs). Resolving the native npm before
+// variants are computed makes reasoning variants produce payloads the native
+// SDKs understand (e.g. anthropic `effort` instead of compat `reasoningEffort`).
+function cloudflareGatewayNpm(providerID: string, modelID: string) {
+  if (providerID !== "cloudflare-ai-gateway") return undefined
+  if (modelID.startsWith("openai/")) return "@ai-sdk/openai"
+  if (modelID.startsWith("anthropic/")) return "@ai-sdk/anthropic"
+  return undefined
+}
+
 function fromModelsDevModel(provider: ModelsDev.Provider, model: ModelsDev.Model): Model {
  const base: Model = {
    id: ModelV2.ID.make(model.id),
@@ -1218,7 +1242,11 @@ function fromModelsDevModel(provider: ModelsDev.Provider, model: ModelsDev.Model
    api: {
      id: model.id,
      url: model.provider?.api ?? provider.api ?? "",
-      npm: model.provider?.npm ?? provider.npm ?? "@ai-sdk/openai-compatible",
+      npm:
+        cloudflareGatewayNpm(provider.id, model.id) ??
+        model.provider?.npm ??
+        provider.npm ??
+        "@ai-sdk/openai-compatible",
    },
    status: model.status ?? "active",
    headers: {},
@@ -1440,6 +1468,9 @@ const layer = Layer.effect(
              model.provider?.npm ??
              provider.npm ??
              existingModel?.api.npm ??
+              // Config-defined gateway models bypass fromModelsDevModel, so resolve the
+              // native passthrough npm here before falling back to the catalog default.
+              cloudflareGatewayNpm(providerID, apiID) ??
              modelsDev[providerID]?.npm ??
              "@ai-sdk/openai-compatible"
            const name = iife(() => {
@@ -1619,6 +1650,7 @@ const layer = Layer.effect(

          for (const [modelID, model] of Object.entries(provider.models)) {
            model.api.id = model.api.id ?? model.id ?? modelID
+
            if (
              // These chat aliases are invalid for the special handling in the
              // built-in providers below, but custom providers may support them.
@@ -56,12 +56,7 @@ import { PermissionV1 } from "@opencode-ai/core/v1/permission"
 import { McpCatalog } from "@/mcp/catalog"

 export function webSearchEnabled(providerID: ProviderV2.ID, flags = { exa: false, parallel: false }) {
-  return (
-    providerID === ProviderV2.ID.opencode ||
-    providerID === ProviderV2.ID.make("opencode-go") ||
-    flags.exa ||
-    flags.parallel
-  )
+  return providerID === ProviderV2.ID.opencode || flags.exa || flags.parallel
 }

 type TaskDef = Tool.InferDef<typeof TaskTool>
@@ -13,56 +13,13 @@ const pluginInput = {
  $: {} as never,
 }

-function makeHookInput(overrides: { providerID?: string; apiId?: string; reasoning?: boolean }) {
-  return {
-    sessionID: "s",
-    agent: "a",
-    provider: {} as never,
-    message: {} as never,
-    model: {
-      providerID: overrides.providerID ?? "cloudflare-ai-gateway",
-      api: { id: overrides.apiId ?? "openai/gpt-5.2-codex", url: "", npm: "ai-gateway-provider" },
-      capabilities: {
-        reasoning: overrides.reasoning ?? true,
-        temperature: false,
-        attachment: true,
-        toolcall: true,
-        input: { text: true, audio: false, image: false, video: false, pdf: false },
-        output: { text: true, audio: false, image: false, video: false, pdf: false },
-        interleaved: false,
-      },
-    } as never,
-  }
-}
-
-function makeHookOutput() {
-  return { temperature: 0, topP: 1, topK: 0, maxOutputTokens: 32_000 as number | undefined, options: {} }
-}
-
-test("omits maxOutputTokens for openai reasoning models on cloudflare-ai-gateway", async () => {
+test("registers the cloudflare-ai-gateway auth method", async () => {
  const hooks = await CloudflareAIGatewayAuthPlugin(pluginInput)
-  const out = makeHookOutput()
-  await hooks["chat.params"]!(makeHookInput({ apiId: "openai/gpt-5.2-codex", reasoning: true }), out)
-  expect(out.maxOutputTokens).toBeUndefined()
+  expect(hooks.auth?.provider).toBe("cloudflare-ai-gateway")
+  expect(hooks.auth?.methods).toHaveLength(1)
 })

-test("keeps maxOutputTokens for openai non-reasoning models", async () => {
+test("no longer drops maxOutputTokens; OpenAI models ride the Responses API passthrough", async () => {
  const hooks = await CloudflareAIGatewayAuthPlugin(pluginInput)
-  const out = makeHookOutput()
-  await hooks["chat.params"]!(makeHookInput({ apiId: "openai/gpt-4-turbo", reasoning: false }), out)
-  expect(out.maxOutputTokens).toBe(32_000)
-})
-
-test("keeps maxOutputTokens for non-openai reasoning models on cloudflare-ai-gateway", async () => {
-  const hooks = await CloudflareAIGatewayAuthPlugin(pluginInput)
-  const out = makeHookOutput()
-  await hooks["chat.params"]!(makeHookInput({ apiId: "anthropic/claude-sonnet-4-5", reasoning: true }), out)
-  expect(out.maxOutputTokens).toBe(32_000)
-})
-
-test("ignores non-cloudflare-ai-gateway providers", async () => {
-  const hooks = await CloudflareAIGatewayAuthPlugin(pluginInput)
-  const out = makeHookOutput()
-  await hooks["chat.params"]!(makeHookInput({ providerID: "openai", apiId: "gpt-5.2-codex", reasoning: true }), out)
-  expect(out.maxOutputTokens).toBe(32_000)
+  expect(hooks["chat.params"]).toBeUndefined()
 })
@@ -1,22 +1,26 @@
-// End-to-end regression test for opencode#24432.
+// End-to-end regression tests for opencode#24432 and opencode#32051/#32052.
 //
-// Routes through the actual ai-gateway-provider + @ai-sdk/openai-compatible
-// chain that provider.ts:811 builds at runtime, with only the network boundary
-// stubbed. Asserts that `reasoning_effort` (and other provider options the
-// transform emits) actually land in the body Cloudflare AI Gateway forwards
-// upstream, which is the only place the bug was observable.
+// Routes through the actual ai-gateway-provider chain that provider.ts builds at
+// runtime, with only the network boundary stubbed:
+//   - openai/*    -> native OpenAI passthrough (Responses API)
+//   - anthropic/* -> native Anthropic passthrough (Messages API)
+//   - everything else -> unified /compat (openai-compatible chat completions)
+// Asserts what actually lands in the envelope body Cloudflare AI Gateway
+// forwards upstream, which is the only place these bugs were observable.

 import { afterEach, beforeEach, describe, expect, test } from "bun:test"
 import type { JSONValue } from "ai"
 import { generateText } from "ai"
 import { createAiGateway } from "ai-gateway-provider"
 import { createUnified } from "ai-gateway-provider/providers/unified"
+import { createOpenAI } from "ai-gateway-provider/providers/openai"
+import { createAnthropic } from "ai-gateway-provider/providers/anthropic"
 import { ProviderTransform } from "@/provider/transform"
 import type * as Provider from "@/provider/provider"
 import { ProviderV2 } from "@opencode-ai/core/provider"
 import { ModelV2 } from "@opencode-ai/core/model"

-type Captured = { url: string; outerBody: unknown }
+type Captured = { url: string; outerBody: unknown; headers: Record<string, string> }
 type ProviderOptions = Record<string, Record<string, JSONValue>>

 const realFetch = globalThis.fetch
@@ -26,24 +30,76 @@ function isRecord(value: unknown): value is Record<string, unknown> {
  return typeof value === "object" && value !== null && !Array.isArray(value)
 }

+// The gateway returns the upstream provider's response body verbatim, so the
+// mock must answer in the wire format of the step's target provider.
+function upstreamResponseBody(provider: string | undefined) {
+  if (provider === "openai")
+    return {
+      id: "resp_test",
+      object: "response",
+      created_at: 0,
+      model: "gpt-5.4",
+      status: "completed",
+      error: null,
+      incomplete_details: null,
+      output: [
+        {
+          type: "message",
+          role: "assistant",
+          id: "msg_1",
+          status: "completed",
+          content: [{ type: "output_text", text: "ok", annotations: [] }],
+        },
+      ],
+      usage: {
+        input_tokens: 1,
+        input_tokens_details: { cached_tokens: 0 },
+        output_tokens: 1,
+        output_tokens_details: { reasoning_tokens: 0 },
+        total_tokens: 2,
+      },
+    }
+  if (provider === "anthropic")
+    return {
+      id: "msg_test",
+      type: "message",
+      role: "assistant",
+      model: "claude-sonnet-4-6",
+      content: [{ type: "text", text: "ok" }],
+      stop_reason: "end_turn",
+      stop_sequence: null,
+      usage: { input_tokens: 1, output_tokens: 1 },
+    }
+  return {
+    id: "chatcmpl-test",
+    object: "chat.completion",
+    created: 0,
+    model: "test",
+    choices: [{ index: 0, message: { role: "assistant", content: "ok" }, finish_reason: "stop" }],
+    usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 },
+  }
+}
+
 beforeEach(() => {
  captured = null
  const handle = async (input: Parameters<typeof fetch>[0], init?: Parameters<typeof fetch>[1]): Promise<Response> => {
    const url = typeof input === "string" ? input : input instanceof URL ? input.toString() : input.url
    if (url.startsWith("https://gateway.ai.cloudflare.com/")) {
      const bodyText = typeof init?.body === "string" ? init.body : ""
-      captured = { url, outerBody: bodyText ? JSON.parse(bodyText) : null }
-      return new Response(
-        JSON.stringify({
-          id: "chatcmpl-test",
-          object: "chat.completion",
-          created: 0,
-          model: "openai/gpt-5.4",
-          choices: [{ index: 0, message: { role: "assistant", content: "ok" }, finish_reason: "stop" }],
-          usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 },
-        }),
-        { status: 200, headers: { "Content-Type": "application/json" } },
-      )
+      const outerBody = bodyText ? JSON.parse(bodyText) : null
+      captured = {
+        url,
+        outerBody,
+        headers: Object.fromEntries(new Headers(init?.headers).entries()),
+      }
+      const provider =
+        Array.isArray(outerBody) && isRecord(outerBody[0]) && typeof outerBody[0].provider === "string"
+          ? outerBody[0].provider
+          : undefined
+      return new Response(JSON.stringify(upstreamResponseBody(provider)), {
+        status: 200,
+        headers: { "Content-Type": "application/json" },
+      })
    }
    return realFetch(input, init)
  }
@@ -56,11 +112,19 @@ afterEach(() => {
  globalThis.fetch = realFetch
 })

+// Mirrors the runtime npm rewrite in provider.ts: openai/anthropic models carry
+// their native SDK package so transforms key provider options correctly.
+const cfNpm = (apiId: string) => {
+  if (apiId.startsWith("openai/")) return "@ai-sdk/openai"
+  if (apiId.startsWith("anthropic/")) return "@ai-sdk/anthropic"
+  return "ai-gateway-provider"
+}
+
 const cfModel = (apiId: string, releaseDate = "2026-03-05"): Provider.Model => ({
  id: ModelV2.ID.make(`cloudflare-ai-gateway/${apiId}`),
  providerID: ProviderV2.ID.make("cloudflare-ai-gateway"),
  name: apiId,
-  api: { id: apiId, url: "https://gateway.ai.cloudflare.com/v1/compat", npm: "ai-gateway-provider" },
+  api: { id: apiId, url: "https://gateway.ai.cloudflare.com/v1/compat", npm: cfNpm(apiId) },
  capabilities: {
    reasoning: true,
    temperature: false,
@@ -80,53 +144,165 @@ const cfModel = (apiId: string, releaseDate = "2026-03-05"): Provider.Model => (

 // ai-gateway-provider sends an array of step descriptors; each entry's `query`
 // is the body forwarded to the upstream provider.
-function extractUpstreamQuery(body: unknown): Record<string, unknown> | undefined {
+function firstStep(body: unknown): Record<string, unknown> | undefined {
  if (!Array.isArray(body) || body.length === 0) return undefined
  const first = body[0]
-  if (!isRecord(first)) return undefined
-  const query = first.query
+  return isRecord(first) ? first : undefined
+}
+
+function extractUpstreamQuery(body: unknown): Record<string, unknown> | undefined {
+  const query = firstStep(body)?.query
  return isRecord(query) ? query : undefined
 }

-async function callThroughGateway(apiId: string, providerOptions: ProviderOptions) {
-  const aigateway = createAiGateway({ accountId: "test", gateway: "test", apiKey: "test" })
-  const unified = createUnified()
-  await generateText({ model: aigateway(unified(apiId)), prompt: "hi", providerOptions })
+// Each step descriptor also carries the `headers` forwarded to the upstream provider.
+function extractUpstreamHeaders(body: unknown): Record<string, unknown> | undefined {
+  const headers = firstStep(body)?.headers
+  return isRecord(headers) ? headers : undefined
+}
+
+// Mirrors the runtime routing in provider.ts getModel.
+function gatewayModel(apiId: string, gatewayToken = "test") {
+  const aigateway = createAiGateway({ accountId: "test", gateway: "test", apiKey: gatewayToken })
+  if (apiId.startsWith("openai/")) return aigateway(createOpenAI()(apiId.slice("openai/".length)))
+  if (apiId.startsWith("anthropic/")) return aigateway(createAnthropic()(apiId.slice("anthropic/".length)))
+  const isWorkersAi = apiId.startsWith("workers-ai/") || apiId.startsWith("@cf/")
+  const unified = createUnified(isWorkersAi ? { apiKey: gatewayToken } : {})
+  return aigateway(unified(apiId))
+}
+
+async function callThroughGateway(apiId: string, providerOptions: ProviderOptions, gatewayToken = "test") {
+  await generateText({ model: gatewayModel(apiId, gatewayToken), prompt: "hi", providerOptions })
  return extractUpstreamQuery(captured?.outerBody)
 }

+describe("cf-ai-gateway routing", () => {
+  test("openai/* rides the native OpenAI passthrough on the Responses API", async () => {
+    await callThroughGateway("openai/gpt-5.4", {})
+    const step = firstStep(captured?.outerBody)
+    expect(step?.provider).toBe("openai")
+    expect(step?.endpoint).toBe("v1/responses")
+    const upstream = extractUpstreamQuery(captured?.outerBody)
+    expect(upstream?.model).toBe("gpt-5.4")
+  })
+
+  test("anthropic/* rides the native Anthropic passthrough on the Messages API", async () => {
+    await callThroughGateway("anthropic/claude-sonnet-4-6", {})
+    const step = firstStep(captured?.outerBody)
+    expect(step?.provider).toBe("anthropic")
+    expect(step?.endpoint).toBe("v1/messages")
+    const upstream = extractUpstreamQuery(captured?.outerBody)
+    expect(upstream?.model).toBe("claude-sonnet-4-6")
+  })
+
+  test("workers-ai models stay on the unified /compat route", async () => {
+    await callThroughGateway("workers-ai/@cf/moonshotai/kimi-k2.6", {})
+    const step = firstStep(captured?.outerBody)
+    expect(step?.provider).toBe("compat")
+    expect(step?.endpoint).toBe("chat/completions")
+    const upstream = extractUpstreamQuery(captured?.outerBody)
+    expect(upstream?.model).toBe("workers-ai/@cf/moonshotai/kimi-k2.6")
+  })
+})
+
 describe("cf-ai-gateway end-to-end (regression: #24432)", () => {
-  test("ProviderTransform.providerOptions output puts reasoning_effort on the wire", async () => {
-    // The full chain the runtime exercises:
-    //   transform.providerOptions() -> openaiCompatible key
-    //   -> @ai-sdk/openai-compatible reads it as compatibleOptions
-    //   -> emits body.reasoning_effort
+  test("ProviderTransform.providerOptions output puts reasoning effort on the Responses wire", async () => {
+    // The full chain the runtime exercises for OpenAI models:
+    //   transform.providerOptions() -> "openai" key (npm rewritten to @ai-sdk/openai)
+    //   -> OpenAIResponsesLanguageModel emits body.reasoning.effort
    //   -> ai-gateway-provider wraps the body and forwards to gateway.ai.cloudflare.com
    const opts = ProviderTransform.providerOptions(cfModel("openai/gpt-5.4"), { reasoningEffort: "xhigh" })
-    expect(opts).toEqual({ openaiCompatible: { reasoningEffort: "xhigh" } })
+    expect(Object.keys(opts)).toEqual(["openai"])
+    expect(opts.openai.reasoningEffort).toBe("xhigh")

    const upstream = await callThroughGateway("openai/gpt-5.4", opts)
-    expect(upstream?.reasoning_effort).toBe("xhigh")
+    expect((upstream?.reasoning as Record<string, unknown> | undefined)?.effort).toBe("xhigh")
  })

  test("variants() output for openai/gpt-5.4 lands xhigh on the wire", async () => {
-    // The other half of the bug: workflow `variant: xhigh` flows through variants()
-    // and must reach the wire. variants() returns the providerOptions payload
-    // unwrapped; providerOptions() wraps it under the SDK key.
+    // fromModelsDevModel resolves the native npm before computing variants, so
+    // OpenAI models get full Responses-flavored payloads (summary + encrypted
+    // reasoning include for stateless multi-turn reasoning).
    const variants = ProviderTransform.variants(cfModel("openai/gpt-5.4"))
-    expect(variants.xhigh).toEqual({ reasoningEffort: "xhigh" })
+    expect(variants.xhigh).toEqual({
+      reasoningEffort: "xhigh",
+      reasoningSummary: "auto",
+      include: ["reasoning.encrypted_content"],
+    })

    const opts = ProviderTransform.providerOptions(cfModel("openai/gpt-5.4"), variants.xhigh)
    const upstream = await callThroughGateway("openai/gpt-5.4", opts)
-    expect(upstream?.reasoning_effort).toBe("xhigh")
+    const reasoning = upstream?.reasoning as Record<string, unknown> | undefined
+    expect(reasoning?.effort).toBe("xhigh")
+    expect(reasoning?.summary).toBe("auto")
+  })
+
+  test("reasoning effort variants for anthropic models land as native adaptive thinking", async () => {
+    // Mirrors the runtime catalog path: models.dev reasoning_options -> reasoningVariants
+    // computed on the native @ai-sdk/anthropic npm -> adaptive thinking + output_config.effort.
+    const model = cfModel("anthropic/claude-sonnet-4-6")
+    const variants = ProviderTransform.reasoningVariants(
+      { reasoning_options: [{ type: "effort", values: ["low", "medium", "high"] }] } as never,
+      model,
+    )
+    expect(variants?.high).toMatchObject({ effort: "high" })
+
+    const opts = ProviderTransform.providerOptions(model, variants!.high)
+    expect(Object.keys(opts)).toEqual(["anthropic"])
+
+    const upstream = await callThroughGateway("anthropic/claude-sonnet-4-6", opts)
+    expect((upstream?.thinking as Record<string, unknown> | undefined)?.type).toBe("adaptive")
+    expect((upstream?.output_config as Record<string, unknown> | undefined)?.effort).toBe("high")
+  })
+
+  test("reasoning_effort still reaches the /compat wire for workers-ai models", async () => {
+    const model = cfModel("workers-ai/@cf/moonshotai/kimi-k2.6")
+    const opts = ProviderTransform.providerOptions(model, { reasoningEffort: "high" })
+    expect(opts).toEqual({ openaiCompatible: { reasoningEffort: "high" } })
+
+    const upstream = await callThroughGateway("workers-ai/@cf/moonshotai/kimi-k2.6", opts)
+    expect(upstream?.reasoning_effort).toBe("high")
  })

  test("legacy buggy key 'cloudflare-ai-gateway' does NOT reach the wire (proves the bug)", async () => {
    // Sanity: confirms the bug class. If a future change accidentally restores
    // providerID-keyed providerOptions, this test fails before users notice.
-    const upstream = await callThroughGateway("openai/gpt-5.4", {
+    const upstream = await callThroughGateway("workers-ai/@cf/moonshotai/kimi-k2.6", {
      "cloudflare-ai-gateway": { reasoningEffort: "high" },
    })
    expect(upstream?.reasoning_effort).toBeUndefined()
  })
 })
+
+describe("cf-ai-gateway token scoping (regression: #32051/#32052)", () => {
+  test("openai passthrough does NOT forward the Cloudflare token upstream", async () => {
+    await callThroughGateway("openai/gpt-5.4", {}, "cf-gateway-secret")
+
+    expect(captured?.headers["cf-aig-authorization"]).toBe("Bearer cf-gateway-secret")
+    // Security invariant: the Cloudflare token must never become the upstream provider's Authorization.
+    expect(extractUpstreamHeaders(captured?.outerBody)?.["authorization"]).toBeUndefined()
+    expect(JSON.stringify(captured?.outerBody)).not.toContain("cf-gateway-secret")
+  })
+
+  test("anthropic passthrough does NOT forward the Cloudflare token upstream", async () => {
+    await callThroughGateway("anthropic/claude-sonnet-4-6", {}, "cf-gateway-secret")
+
+    expect(captured?.headers["cf-aig-authorization"]).toBe("Bearer cf-gateway-secret")
+    expect(extractUpstreamHeaders(captured?.outerBody)?.["x-api-key"]).toBeUndefined()
+    expect(JSON.stringify(captured?.outerBody)).not.toContain("cf-gateway-secret")
+  })
+
+  test("workers-ai models DO forward the Cloudflare token upstream", async () => {
+    await callThroughGateway("workers-ai/@cf/google/gemma-4-26b-a4b-it", {}, "cf-gateway-secret")
+
+    expect(captured?.headers["cf-aig-authorization"]).toBe("Bearer cf-gateway-secret")
+    expect(extractUpstreamHeaders(captured?.outerBody)?.["authorization"]).toBe("Bearer cf-gateway-secret")
+  })
+
+  test("bare @cf/ Workers AI models DO forward the Cloudflare token upstream", async () => {
+    await callThroughGateway("@cf/meta/llama-3.1-8b-instruct", {}, "cf-gateway-secret")
+
+    expect(captured?.headers["cf-aig-authorization"]).toBe("Bearer cf-gateway-secret")
+    expect(extractUpstreamHeaders(captured?.outerBody)?.["authorization"]).toBe("Bearer cf-gateway-secret")
+  })
+})
@@ -37,9 +37,8 @@ describe("websearch provider", () => {
    expect(selectWebSearchProvider(SESSION_ID, { exa: false, parallel: true })).toBe("parallel")
  })

-  test("is enabled for OpenCode providers or explicit websearch provider flags", () => {
+  test("is only enabled for opencode or explicit websearch provider flags", () => {
    expect(webSearchEnabled(ProviderV2.ID.opencode, { exa: false, parallel: false })).toBe(true)
-    expect(webSearchEnabled(ProviderV2.ID.make("opencode-go"), { exa: false, parallel: false })).toBe(true)
    expect(webSearchEnabled(ProviderV2.ID.openai, { exa: false, parallel: false })).toBe(false)
    expect(webSearchEnabled(ProviderV2.ID.openai, { exa: true, parallel: false })).toBe(true)
    expect(webSearchEnabled(ProviderV2.ID.openai, { exa: false, parallel: true })).toBe(true)
@@ -50,7 +50,6 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال
 تشمل قائمة النماذج الحالية:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال
 | ----------------- | ------------------- | ------------------ | ---------------- |
 | Grok 4.5          | 120                 | 300                | 600              |
 | GPT 5.6 Luna      | 2,050               | 5,100              | 10,250           |
-| GLM-5.3           | 220                 | 540                | 1,080            |
 | GLM-5.2           | 880                 | 2,150              | 4,300            |
 | GLM-5.1           | 880                 | 2,150              | 4,300            |
 | Kimi K3           | 110                 | 250                | 490              |
@@ -110,7 +108,7 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال
 تستند التقديرات إلى أنماط الطلبات المرصودة:

 - Grok 4.5 — ‏1,100 input، و71,500 cached، و220 output tokens لكل طلب
- GLM-5.3/5.2/5.1 — ‏700 input، و52,000 cached، و150 output tokens لكل طلب
+- GLM-5.2/5.1 — ‏700 input، و52,000 cached، و150 output tokens لكل طلب
 - GPT 5.6 Luna — ‏1,000 توكن إدخال، و50,000 توكن مخزّن مؤقتًا، و220 توكن إخراج لكل طلب
 - Kimi K3 — ‏1,050 input، و76,500 cached، و300 output tokens لكل طلب
 - Kimi K2.7/K2.6 — ‏870 input، و55,000 cached، و200 output tokens لكل طلب
@@ -133,7 +131,6 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال
 | Grok 4.5                     | $2.00   | $6.00   | $0.30           | -               | $15       |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20   | $1.20   | $0.02           | $0.25           | $15       |
 | GPT 5.6 Luna (> 272K tokens) | $0.40   | $1.80   | $0.04           | $0.50           | $15       |
-| GLM-5.3                      | $1.40   | $4.40   | $0.26           | -               | $15       |
 | GLM-5.2                      | $1.40   | $4.40   | $0.26           | -               | $60       |
 | GLM-5.1                      | $1.40   | $4.40   | $0.26           | -               | $60       |
 | Kimi K3                      | $3.00   | $15.00  | $0.30           | -               | $15       |
@@ -190,7 +187,6 @@ OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر ال
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------- | ------------------ |
 | Grok 4.5          | غير مستخدَمة  | 30 يومًا           |
 | GPT 5.6 Luna      | غير مستخدَمة  | 30 يومًا           |
-| GLM-5.3           | غير مستخدَمة  | 0 أيام             |
 | GLM-5.2           | غير مستخدَمة  | 0 أيام             |
 | GLM-5.1           | غير مستخدَمة  | 0 أيام             |
 | Kimi K3           | غير مستخدَمة  | 0 أيام             |
@@ -94,7 +94,6 @@ OpenCode Zen هي بوابة AI تتيح لك الوصول إلى هذه الن
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -185,7 +184,6 @@ https://opencode.ai/zen/v1/models
 | 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           | -               |
-| Muse Spark 1.2                    | $1.25   | $4.25   | $0.15           | -               |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00   | $30.00  | $0.50           | $6.25           |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00  | $45.00  | $1.00           | $12.50          |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00   | $12.00  | $0.20           | $2.50           |
@@ -60,7 +60,6 @@ Samo jedan član po radnom prostoru (workspace) može se pretplatiti na OpenCode
 Trenutna lista modela uključuje:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -99,7 +98,6 @@ Tabela ispod pruža procijenjeni broj zahtjeva na osnovu tipičnih obrazaca kori
 | ----------------- | ------------------ | ----------------- | ----------------- |
 | Grok 4.5          | 120                | 300               | 600               |
 | GPT 5.6 Luna      | 2,050              | 5,100             | 10,250            |
-| GLM-5.3           | 220                | 540               | 1,080             |
 | GLM-5.2           | 880                | 2,150             | 4,300             |
 | GLM-5.1           | 880                | 2,150             | 4,300             |
 | Kimi K3           | 110                | 250               | 490               |
@@ -120,7 +118,7 @@ Tabela ispod pruža procijenjeni broj zahtjeva na osnovu tipičnih obrazaca kori
 Procjene se zasnivaju na zapaženim obrascima zahtjeva:

 - Grok 4.5 — 1,100 ulaznih, 71,500 keširanih, 220 izlaznih tokena po zahtjevu
- GLM-5.3/5.2/5.1 — 700 ulaznih (input), 52,000 keširanih, 150 izlaznih (output) tokena po zahtjevu
+- GLM-5.2/5.1 — 700 ulaznih (input), 52,000 keširanih, 150 izlaznih (output) tokena po zahtjevu
 - GPT 5.6 Luna — 1,000 ulaznih, 50,000 keširanih, 220 izlaznih tokena po zahtjevu
 - Kimi K3 — 1,050 ulaznih, 76,500 keširanih, 300 izlaznih tokena po zahtjevu
 - Kimi K2.7/K2.6 — 870 ulaznih, 55,000 keširanih, 200 izlaznih tokena po zahtjevu
@@ -143,7 +141,6 @@ Procjene se također zasnivaju na sljedećim cijenama po 1M tokena i mjesečnoj
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15       |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15       |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15       |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15       |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60       |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60       |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15       |
@@ -202,7 +199,6 @@ Također možete pristupiti Go modelima putem sljedećih API endpointa.
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -243,7 +239,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ----------------- | -------------------- |
 | Grok 4.5          | Ne koristi se     | 30 dana              |
 | GPT 5.6 Luna      | Ne koristi se     | 30 dana              |
-| GLM-5.3           | Ne koristi se     | 0 dana               |
 | GLM-5.2           | Ne koristi se     | 0 dana               |
 | GLM-5.1           | Ne koristi se     | 0 dana               |
 | Kimi K3           | Ne koristi se     | 0 dana               |
@@ -99,7 +99,6 @@ Našim modelima možete pristupiti i preko sljedećih API endpointa.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ Podržavamo pay-as-you-go model. Ispod su cijene **po 1M tokena**.
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -60,7 +60,6 @@ Kun ét medlem per arbejdsområde kan abonnere på OpenCode Go.
 Den nuværende liste over modeller inkluderer:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -99,7 +98,6 @@ Tabellen nedenfor giver et estimeret antal anmodninger baseret på typiske Go-fo
 | ----------------- | ----------------------- | ------------------- | --------------------- |
 | Grok 4.5          | 120                     | 300                 | 600                   |
 | GPT 5.6 Luna      | 2,050                   | 5,100               | 10,250                |
-| GLM-5.3           | 220                     | 540                 | 1,080                 |
 | GLM-5.2           | 880                     | 2,150               | 4,300                 |
 | GLM-5.1           | 880                     | 2,150               | 4,300                 |
 | Kimi K3           | 110                     | 250                 | 490                   |
@@ -120,7 +118,7 @@ Tabellen nedenfor giver et estimeret antal anmodninger baseret på typiske Go-fo
 Estimaterne er baseret på observerede anmodningsmønstre:

 - Grok 4.5 — 1.100 input, 71.500 cachelagrede, 220 output-tokens pr. anmodning
- GLM-5.3/5.2/5.1 — 700 input, 52.000 cachelagrede, 150 output-tokens pr. anmodning
+- GLM-5.2/5.1 — 700 input, 52.000 cachelagrede, 150 output-tokens pr. anmodning
 - GPT 5.6 Luna — 1.000 input, 50.000 cachelagrede, 220 output-tokens pr. anmodning
 - Kimi K3 — 1.050 input, 76.500 cachelagrede, 300 output-tokens pr. anmodning
 - Kimi K2.7/K2.6 — 870 input, 55.000 cachelagrede, 200 output-tokens pr. anmodning
@@ -143,7 +141,6 @@ Estimaterne er også baseret på følgende priser pr. 1M tokens og det månedlig
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15     |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15     |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15     |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15     |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60     |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60     |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15     |
@@ -202,7 +199,6 @@ Du kan også få adgang til Go-modeller gennem følgende API-endpoints.
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -243,7 +239,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------ | -------------- |
 | Grok 4.5          | Ikke brugt   | 30 dage        |
 | GPT 5.6 Luna      | Ikke brugt   | 30 dage        |
-| GLM-5.3           | Ikke brugt   | 0 dage         |
 | GLM-5.2           | Ikke brugt   | 0 dage         |
 | GLM-5.1           | Ikke brugt   | 0 dage         |
 | Kimi K3           | Ikke brugt   | 0 dage         |
@@ -99,7 +99,6 @@ Du kan også få adgang til vores modeller gennem følgende API-endpoints.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ Vi understøtter en pay-as-you-go-model. Nedenfor er priserne **pr. 1M tokens**.
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -52,7 +52,6 @@ Nur ein Mitglied pro Workspace kann OpenCode Go abonnieren.
 Die aktuelle Liste der Modelle umfasst:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -91,7 +90,6 @@ Die folgende Tabelle zeigt eine geschätzte Anzahl von Anfragen basierend auf ty
 | ----------------- | ---------------------- | ------------------ | ------------------ |
 | Grok 4.5          | 120                    | 300                | 600                |
 | GPT 5.6 Luna      | 2,050                  | 5,100              | 10,250             |
-| GLM-5.3           | 220                    | 540                | 1,080              |
 | GLM-5.2           | 880                    | 2,150              | 4,300              |
 | GLM-5.1           | 880                    | 2,150              | 4,300              |
 | Kimi K3           | 110                    | 250                | 490                |
@@ -112,7 +110,7 @@ Die folgende Tabelle zeigt eine geschätzte Anzahl von Anfragen basierend auf ty
 Die Schätzungen basieren auf beobachteten Anfragemustern:

 - Grok 4.5 — 1.100 Input-, 71.500 Cached-, 220 Output-Tokens pro Anfrage
- GLM-5.3/5.2/5.1 — 700 Input-, 52.000 Cached-, 150 Output-Tokens pro Anfrage
+- GLM-5.2/5.1 — 700 Input-, 52.000 Cached-, 150 Output-Tokens pro Anfrage
 - GPT 5.6 Luna — 1.000 Input-, 50.000 Cached-, 220 Output-Tokens pro Anfrage
 - Kimi K3 — 1.050 Input-, 76.500 Cached-, 300 Output-Tokens pro Anfrage
 - Kimi K2.7/K2.6 — 870 Input-, 55.000 Cached-, 200 Output-Tokens pro Anfrage
@@ -135,7 +133,6 @@ Die Schätzungen basieren außerdem auf den folgenden Preisen pro 1M Tokens und
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15     |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15     |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15     |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15     |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60     |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60     |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15     |
@@ -192,7 +189,6 @@ Du kannst auf die Go-Modelle auch über die folgenden API-Endpunkte zugreifen.
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -231,7 +227,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | --------------- | ----------------- |
 | Grok 4.5          | Nicht verwendet | 30 Tage           |
 | GPT 5.6 Luna      | Nicht verwendet | 30 Tage           |
-| GLM-5.3           | Nicht verwendet | 0 Tage            |
 | GLM-5.2           | Nicht verwendet | 0 Tage            |
 | GLM-5.1           | Nicht verwendet | 0 Tage            |
 | Kimi K3           | Nicht verwendet | 0 Tage            |
@@ -90,7 +90,6 @@ Du kannst auch über die folgenden API-Endpunkte auf unsere Modelle zugreifen.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -181,7 +180,6 @@ Wir unterstützen ein Pay-as-you-go-Modell. Unten findest du die Preise **pro 1M
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -60,7 +60,6 @@ Solo un miembro por espacio de trabajo puede suscribirse a OpenCode Go.
 La lista actual de modelos incluye:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -99,7 +98,6 @@ La siguiente tabla proporciona una cantidad estimada de peticiones basada en los
 | ----------------- | ---------------------- | --------------------- | ------------------ |
 | Grok 4.5          | 120                    | 300                   | 600                |
 | GPT 5.6 Luna      | 2,050                  | 5,100                 | 10,250             |
-| GLM-5.3           | 220                    | 540                   | 1,080              |
 | GLM-5.2           | 880                    | 2,150                 | 4,300              |
 | GLM-5.1           | 880                    | 2,150                 | 4,300              |
 | Kimi K3           | 110                    | 250                   | 490                |
@@ -120,7 +118,7 @@ La siguiente tabla proporciona una cantidad estimada de peticiones basada en los
 Las estimaciones se basan en los patrones de peticiones observados:

 - Grok 4.5 — 1,100 tokens de entrada, 71,500 en caché, 220 tokens de salida por petición
- GLM-5.3/5.2/5.1 — 700 tokens de entrada, 52,000 en caché, 150 tokens de salida por petición
+- GLM-5.2/5.1 — 700 tokens de entrada, 52,000 en caché, 150 tokens de salida por petición
 - GPT 5.6 Luna — 1,000 tokens de entrada, 50,000 en caché, 220 tokens de salida por petición
 - Kimi K3 — 1,050 tokens de entrada, 76,500 en caché, 300 tokens de salida por petición
 - Kimi K2.7/K2.6 — 870 tokens de entrada, 55,000 en caché, 200 tokens de salida por petición
@@ -143,7 +141,6 @@ Las estimaciones también se basan en los siguientes precios por 1M tokens y en
 | Grok 4.5                     | $2.00   | $6.00  | $0.30            | -                  | $15 |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20   | $1.20  | $0.02            | $0.25              | $15 |
 | GPT 5.6 Luna (> 272K tokens) | $0.40   | $1.80  | $0.04            | $0.50              | $15 |
-| GLM-5.3                      | $1.40   | $4.40  | $0.26            | -                  | $15 |
 | GLM-5.2                      | $1.40   | $4.40  | $0.26            | -                  | $60 |
 | GLM-5.1                      | $1.40   | $4.40  | $0.26            | -                  | $60 |
 | Kimi K3                      | $3.00   | $15.00 | $0.30            | -                  | $15 |
@@ -202,7 +199,6 @@ También puedes acceder a los modelos de Go a través de los siguientes endpoint
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -243,7 +239,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------------------ | ------------------ |
 | Grok 4.5          | No utilizado             | 30 días            |
 | GPT 5.6 Luna      | No utilizado             | 30 días            |
-| GLM-5.3           | No utilizado             | 0 días             |
 | GLM-5.2           | No utilizado             | 0 días             |
 | GLM-5.1           | No utilizado             | 0 días             |
 | Kimi K3           | No utilizado             | 0 días             |
@@ -99,7 +99,6 @@ También puedes acceder a nuestros modelos a través de los siguientes endpoints
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ Admitimos un modelo de pago por uso. A continuación se muestran los precios **p
 | 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            | -                  |
-| Muse Spark 1.2                    | $1.25   | $4.25   | $0.15            | -                  |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00   | $30.00  | $0.50            | $6.25              |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00  | $45.00  | $1.00            | $12.50             |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00   | $12.00  | $0.20            | $2.50              |
@@ -50,7 +50,6 @@ Un seul membre par espace de travail peut s'abonner à OpenCode Go.
 La liste actuelle des modèles comprend :

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ Le tableau ci-dessous fournit une estimation du nombre de requêtes basée sur d
 | ----------------- | --------------------- | -------------------- | ----------------- |
 | Grok 4.5          | 120                   | 300                  | 600               |
 | GPT 5.6 Luna      | 2,050                 | 5,100                | 10,250            |
-| GLM-5.3           | 220                   | 540                  | 1,080             |
 | GLM-5.2           | 880                   | 2,150                | 4,300             |
 | GLM-5.1           | 880                   | 2,150                | 4,300             |
 | Kimi K3           | 110                   | 250                  | 490               |
@@ -110,7 +108,7 @@ Le tableau ci-dessous fournit une estimation du nombre de requêtes basée sur d
 Les estimations sont basées sur les schémas de requêtes observés :

 - Grok 4.5 — 1,100 tokens en entrée, 71,500 en cache, 220 tokens en sortie par requête
- GLM-5.3/5.2/5.1 — 700 tokens en entrée, 52,000 en cache, 150 tokens en sortie par requête
+- GLM-5.2/5.1 — 700 tokens en entrée, 52,000 en cache, 150 tokens en sortie par requête
 - GPT 5.6 Luna — 1,000 tokens en entrée, 50,000 en cache, 220 tokens en sortie par requête
 - Kimi K3 — 1,050 tokens en entrée, 76,500 en cache, 300 tokens en sortie par requête
 - Kimi K2.7/K2.6 — 870 tokens en entrée, 55,000 en cache, 200 tokens en sortie par requête
@@ -133,7 +131,6 @@ Les estimations sont également basées sur les prix suivants par 1M tokens et s
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15         |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15         |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15         |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15         |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60         |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60         |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15         |
@@ -190,7 +187,6 @@ Vous pouvez également accéder aux modèles Go via les points de terminaison d'
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------------------ | ------------------------ |
 | Grok 4.5          | Non utilisé              | 30 jours                 |
 | GPT 5.6 Luna      | Non utilisé              | 30 jours                 |
-| GLM-5.3           | Non utilisé              | 0 jour                   |
 | GLM-5.2           | Non utilisé              | 0 jour                   |
 | GLM-5.1           | Non utilisé              | 0 jour                   |
 | Kimi K3           | Non utilisé              | 0 jour                   |
@@ -90,7 +90,6 @@ Vous pouvez également accéder à nos modèles via les points de terminaison AP
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -181,7 +180,6 @@ Nous prenons en charge un modèle de paiement à l'utilisation. Vous trouverez c
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -60,7 +60,6 @@ Only one member per workspace can subscribe to OpenCode Go.
 The current list of models includes:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -99,7 +98,6 @@ The table below provides an estimated request count based on typical Go usage pa
 | ----------------- | ------------------- | ----------------- | ------------------ |
 | Grok 4.5          | 120                 | 300               | 600                |
 | GPT 5.6 Luna      | 2,050               | 5,100             | 10,250             |
-| GLM-5.3           | 220                 | 540               | 1,080              |
 | GLM-5.2           | 880                 | 2,150             | 4,300              |
 | GLM-5.1           | 880                 | 2,150             | 4,300              |
 | Kimi K3           | 110                 | 250               | 490                |
@@ -120,7 +118,7 @@ The table below provides an estimated request count based on typical Go usage pa
 The estimates are based on observed request patterns:

 - Grok 4.5 — 1,100 input, 71,500 cached, 220 output tokens per request
- GLM-5.3/5.2/5.1 — 700 input, 52,000 cached, 150 output tokens per request
+- GLM-5.2/5.1 — 700 input, 52,000 cached, 150 output tokens per request
 - GPT 5.6 Luna — 1,000 input, 50,000 cached, 220 output tokens per request
 - Kimi K3 — 1,050 input, 76,500 cached, 300 output tokens per request
 - Kimi K2.7/K2.6 — 870 input, 55,000 cached, 200 output tokens per request
@@ -143,7 +141,6 @@ The estimates are also based on the following prices per 1M tokens and the month
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15   |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15   |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15   |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15   |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15   |
@@ -202,7 +199,6 @@ You can also access Go models through the following API endpoints.
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -243,7 +239,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | -------------- | -------------- |
 | Grok 4.5          | Not used       | 30 days        |
 | GPT 5.6 Luna      | Not used       | 30 days        |
-| GLM-5.3           | Not used       | 0 days         |
 | GLM-5.2           | Not used       | 0 days         |
 | GLM-5.1           | Not used       | 0 days         |
 | Kimi K3           | Not used       | 0 days         |
@@ -58,7 +58,6 @@ Solo un membro per workspace può abbonarsi a OpenCode Go.
 L'elenco attuale dei modelli include:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -97,7 +96,6 @@ La tabella seguente fornisce una stima del conteggio delle richieste in base a p
 | ----------------- | -------------------- | --------------------- | ----------------- |
 | Grok 4.5          | 120                  | 300                   | 600               |
 | GPT 5.6 Luna      | 2,050                | 5,100                 | 10,250            |
-| GLM-5.3           | 220                  | 540                   | 1,080             |
 | GLM-5.2           | 880                  | 2,150                 | 4,300             |
 | GLM-5.1           | 880                  | 2,150                 | 4,300             |
 | Kimi K3           | 110                  | 250                   | 490               |
@@ -118,7 +116,7 @@ La tabella seguente fornisce una stima del conteggio delle richieste in base a p
 Le stime si basano sui pattern di richieste osservati:

 - Grok 4.5 — 1.100 di input, 71.500 in cache, 220 token di output per richiesta
- GLM-5.3/5.2/5.1 — 700 di input, 52.000 in cache, 150 token di output per richiesta
+- GLM-5.2/5.1 — 700 di input, 52.000 in cache, 150 token di output per richiesta
 - GPT 5.6 Luna — 1.000 token di input, 50.000 in cache, 220 token di output per richiesta
 - Kimi K3 — 1.050 di input, 76.500 in cache, 300 token di output per richiesta
 - Kimi K2.7/K2.6 — 870 di input, 55.000 in cache, 200 token di output per richiesta
@@ -141,7 +139,6 @@ Le stime si basano anche sui seguenti prezzi per 1M token e sull'utilizzo mensil
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15      |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15      |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15      |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15      |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60      |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60      |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15      |
@@ -200,7 +197,6 @@ Puoi anche accedere ai modelli Go tramite i seguenti endpoint API.
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -241,7 +237,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------------------- | ---------------------- |
 | Grok 4.5          | Non utilizzato            | 30 giorni              |
 | GPT 5.6 Luna      | Non utilizzato            | 30 giorni              |
-| GLM-5.3           | Non utilizzato            | 0 giorni               |
 | GLM-5.2           | Non utilizzato            | 0 giorni               |
 | GLM-5.1           | Non utilizzato            | 0 giorni               |
 | Kimi K3           | Non utilizzato            | 0 giorni               |
@@ -99,7 +99,6 @@ Puoi anche accedere ai nostri modelli tramite i seguenti endpoint API.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ Supportiamo un modello pay-as-you-go. Qui sotto trovi i prezzi **per 1M token**.
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -50,7 +50,6 @@ OpenCode Goをサブスクライブできるのは、1つのワークスペー
 現在のモデルリストには以下が含まれます：

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ OpenCode Goには以下の制限が含まれています：
 | ----------------- | ------------------------- | ---------------- | ---------------- |
 | Grok 4.5          | 120                       | 300              | 600              |
 | GPT 5.6 Luna      | 2,050                     | 5,100            | 10,250           |
-| GLM-5.3           | 220                       | 540              | 1,080            |
 | GLM-5.2           | 880                       | 2,150            | 4,300            |
 | GLM-5.1           | 880                       | 2,150            | 4,300            |
 | Kimi K3           | 110                       | 250              | 490              |
@@ -110,7 +108,7 @@ OpenCode Goには以下の制限が含まれています：
 推定値は、観測されたリクエストパターンに基づいています：

 - Grok 4.5 — リクエストあたり 入力 1,100トークン、キャッシュ 71,500トークン、出力 220トークン
- GLM-5.3/5.2/5.1 — リクエストあたり 入力 700トークン、キャッシュ 52,000トークン、出力 150トークン
+- GLM-5.2/5.1 — リクエストあたり 入力 700トークン、キャッシュ 52,000トークン、出力 150トークン
 - GPT 5.6 Luna — リクエストあたり 入力 1,000トークン、キャッシュ 50,000トークン、出力 220トークン
 - Kimi K3 — リクエストあたり 入力 1,050トークン、キャッシュ 76,500トークン、出力 300トークン
 - Kimi K2.7/K2.6 — リクエストあたり 入力 870トークン、キャッシュ 55,000トークン、出力 200トークン
@@ -133,7 +131,6 @@ OpenCode Goには以下の制限が含まれています：
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15   |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15   |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15   |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15   |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15   |
@@ -190,7 +187,6 @@ Goでは月額$10を支払い、その6倍の利用枠を提供することを
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | -------------------- | ---------- |
 | Grok 4.5          | 使用なし             | 30日       |
 | GPT 5.6 Luna      | 使用なし             | 30日       |
-| GLM-5.3           | 使用なし             | 0日        |
 | GLM-5.2           | 使用なし             | 0日        |
 | GLM-5.1           | 使用なし             | 0日        |
 | Kimi K3           | 使用なし             | 0日        |
@@ -90,7 +90,6 @@ OpenCode Zen は、OpenCode のほかのプロバイダーと同じように動
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -181,7 +180,6 @@ https://opencode.ai/zen/v1/models
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -50,7 +50,6 @@ workspace당 한 명의 멤버만 OpenCode Go를 구독할 수 있습니다.
 현재 모델 목록에는 다음이 포함됩니다.

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ OpenCode Go에는 다음과 같은 한도가 포함됩니다.
 | ----------------- | ----------------- | -------------- | -------------- |
 | Grok 4.5          | 120               | 300            | 600            |
 | GPT 5.6 Luna      | 2,050             | 5,100          | 10,250         |
-| GLM-5.3           | 220               | 540            | 1,080          |
 | GLM-5.2           | 880               | 2,150          | 4,300          |
 | GLM-5.1           | 880               | 2,150          | 4,300          |
 | Kimi K3           | 110               | 250            | 490            |
@@ -110,7 +108,7 @@ OpenCode Go에는 다음과 같은 한도가 포함됩니다.
 이 예상치는 관찰된 요청 패턴을 기준으로 합니다.

 - Grok 4.5 — 요청당 입력 1,100, 캐시 71,500, 출력 토큰 220
- GLM-5.3/5.2/5.1 — 요청당 입력 700, 캐시 52,000, 출력 토큰 150
+- GLM-5.2/5.1 — 요청당 입력 700, 캐시 52,000, 출력 토큰 150
 - GPT 5.6 Luna — 요청당 입력 토큰 1,000개, 캐시 토큰 50,000개, 출력 토큰 220개
 - Kimi K3 — 요청당 입력 1,050, 캐시 76,500, 출력 토큰 300
 - Kimi K2.7/K2.6 — 요청당 입력 870, 캐시 55,000, 출력 토큰 200
@@ -133,7 +131,6 @@ OpenCode Go에는 다음과 같은 한도가 포함됩니다.
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15   |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15   |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15   |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15   |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15   |
@@ -190,7 +187,6 @@ Go에서는 월 $10를 지불하며, 저희는 그 6배의 사용량을 제공
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------- | ----------- |
 | Grok 4.5          | 사용되지 않음 | 30일        |
 | GPT 5.6 Luna      | 사용되지 않음 | 30일        |
-| GLM-5.3           | 사용되지 않음 | 0일         |
 | GLM-5.2           | 사용되지 않음 | 0일         |
 | GLM-5.1           | 사용되지 않음 | 0일         |
 | Kimi K3           | 사용되지 않음 | 0일         |
@@ -90,7 +90,6 @@ OpenCode Zen은 OpenCode의 다른 provider와 똑같이 작동합니다.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -181,7 +180,6 @@ https://opencode.ai/zen/v1/models
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -60,7 +60,6 @@ Kun ett medlem per arbeidsområde kan abonnere på OpenCode Go.
 Den nåværende listen over modeller inkluderer:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -99,7 +98,6 @@ Tabellen nedenfor gir et estimert antall forespørsler basert på typiske bruksm
 | ----------------- | ------------------------ | -------------------- | ---------------------- |
 | Grok 4.5          | 120                      | 300                  | 600                    |
 | GPT 5.6 Luna      | 2,050                    | 5,100                | 10,250                 |
-| GLM-5.3           | 220                      | 540                  | 1,080                  |
 | GLM-5.2           | 880                      | 2,150                | 4,300                  |
 | GLM-5.1           | 880                      | 2,150                | 4,300                  |
 | Kimi K3           | 110                      | 250                  | 490                    |
@@ -120,7 +118,7 @@ Tabellen nedenfor gir et estimert antall forespørsler basert på typiske bruksm
 Estimatene er basert på observerte forespørselsmønstre:

 - Grok 4.5 — 1 100 input, 71 500 bufret, 220 output-tokens per forespørsel
- GLM-5.3/5.2/5.1 — 700 input, 52 000 bufret, 150 output-tokens per forespørsel
+- GLM-5.2/5.1 — 700 input, 52 000 bufret, 150 output-tokens per forespørsel
 - GPT 5.6 Luna — 1 000 input, 50 000 bufret, 220 output-tokens per forespørsel
 - Kimi K3 — 1 050 input, 76 500 bufret, 300 output-tokens per forespørsel
 - Kimi K2.7/K2.6 — 870 input, 55 000 bufret, 200 output-tokens per forespørsel
@@ -143,7 +141,6 @@ Estimatene er også basert på følgende priser per 1M tokens og den månedlige
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15  |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15  |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15  |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15  |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60  |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60  |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15  |
@@ -202,7 +199,6 @@ Du kan også få tilgang til Go-modeller gjennom følgende API-endepunkter.
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -243,7 +239,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------- | --------------- |
 | Grok 4.5          | Brukes ikke   | 30 dager        |
 | GPT 5.6 Luna      | Brukes ikke   | 30 dager        |
-| GLM-5.3           | Brukes ikke   | 0 dager         |
 | GLM-5.2           | Brukes ikke   | 0 dager         |
 | GLM-5.1           | Brukes ikke   | 0 dager         |
 | Kimi K3           | Brukes ikke   | 0 dager         |
@@ -99,7 +99,6 @@ Du kan også få tilgang til modellene våre gjennom følgende API-endepunkter.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ Vi støtter en pay-as-you-go-modell. Nedenfor er prisene **per 1M tokens**.
 | 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         | -               |
-| Muse Spark 1.2                    | $1.25   | $4.25   | $0.15         | -               |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00   | $30.00  | $0.50         | $6.25           |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00  | $45.00  | $1.00         | $12.50          |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00   | $12.00  | $0.20         | $2.50           |
@@ -54,7 +54,6 @@ Tylko jeden członek na obszar roboczy (workspace) może zasubskrybować OpenCod
 Obecna lista modeli obejmuje:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -93,7 +92,6 @@ Poniższa tabela przedstawia szacunkową liczbę żądań na podstawie typowych
 | ----------------- | ------------------- | ------------------ | ------------------ |
 | Grok 4.5          | 120                 | 300                | 600                |
 | GPT 5.6 Luna      | 2,050               | 5,100              | 10,250             |
-| GLM-5.3           | 220                 | 540                | 1,080              |
 | GLM-5.2           | 880                 | 2,150              | 4,300              |
 | GLM-5.1           | 880                 | 2,150              | 4,300              |
 | Kimi K3           | 110                 | 250                | 490                |
@@ -114,7 +112,7 @@ Poniższa tabela przedstawia szacunkową liczbę żądań na podstawie typowych
 Szacunki te opierają się na zaobserwowanych wzorcach żądań:

 - Grok 4.5 — 1 100 tokenów wejściowych, 71 500 w pamięci podręcznej, 220 tokenów wyjściowych na żądanie
- GLM-5.3/5.2/5.1 — 700 tokenów wejściowych, 52 000 w pamięci podręcznej, 150 tokenów wyjściowych na żądanie
+- GLM-5.2/5.1 — 700 tokenów wejściowych, 52 000 w pamięci podręcznej, 150 tokenów wyjściowych na żądanie
 - GPT 5.6 Luna — 1 000 tokenów wejściowych, 50 000 w pamięci podręcznej, 220 tokenów wyjściowych na żądanie
 - Kimi K3 — 1 050 tokenów wejściowych, 76 500 w pamięci podręcznej, 300 tokenów wyjściowych na żądanie
 - Kimi K2.7/K2.6 — 870 tokenów wejściowych, 55 000 w pamięci podręcznej, 200 tokenów wyjściowych na żądanie
@@ -137,7 +135,6 @@ Szacunki opierają się również na następujących cenach za 1M tokenów oraz
 | Grok 4.5                     | $2.00   | $6.00   | $0.30          | -              | $15    |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20   | $1.20   | $0.02          | $0.25          | $15    |
 | GPT 5.6 Luna (> 272K tokens) | $0.40   | $1.80   | $0.04          | $0.50          | $15    |
-| GLM-5.3                      | $1.40   | $4.40   | $0.26          | -              | $15    |
 | GLM-5.2                      | $1.40   | $4.40   | $0.26          | -              | $60    |
 | GLM-5.1                      | $1.40   | $4.40   | $0.26          | -              | $60    |
 | Kimi K3                      | $3.00   | $15.00  | $0.30          | -              | $15    |
@@ -194,7 +191,6 @@ Możesz również uzyskać dostęp do modeli Go za pośrednictwem następującyc
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -235,7 +231,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ----------------- | --------------- |
 | Grok 4.5          | Niewykorzystywane | 30 dni          |
 | GPT 5.6 Luna      | Niewykorzystywane | 30 dni          |
-| GLM-5.3           | Niewykorzystywane | 0 dni           |
 | GLM-5.2           | Niewykorzystywane | 0 dni           |
 | GLM-5.1           | Niewykorzystywane | 0 dni           |
 | Kimi K3           | Niewykorzystywane | 0 dni           |
@@ -99,7 +99,6 @@ Możesz też uzyskać dostęp do naszych modeli przez poniższe endpointy API.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ Obsługujemy model pay-as-you-go. Poniżej znajdują się ceny **za 1M tokenów*
 | 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          | -              |
-| Muse Spark 1.2                    | $1.25   | $4.25   | $0.15          | -              |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00   | $30.00  | $0.50          | $6.25          |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00  | $45.00  | $1.00          | $12.50         |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00   | $12.00  | $0.20          | $2.50          |
@@ -60,7 +60,6 @@ Apenas um membro por workspace pode assinar o OpenCode Go.
 A lista atual de modelos inclui:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -99,7 +98,6 @@ A tabela abaixo fornece uma contagem estimada de requisições com base nos padr
 | ----------------- | ----------------------- | ---------------------- | ------------------- |
 | Grok 4.5          | 120                     | 300                    | 600                 |
 | GPT 5.6 Luna      | 2,050                   | 5,100                  | 10,250              |
-| GLM-5.3           | 220                     | 540                    | 1,080               |
 | GLM-5.2           | 880                     | 2,150                  | 4,300               |
 | GLM-5.1           | 880                     | 2,150                  | 4,300               |
 | Kimi K3           | 110                     | 250                    | 490                 |
@@ -120,7 +118,7 @@ A tabela abaixo fornece uma contagem estimada de requisições com base nos padr
 As estimativas se baseiam nos padrões de requisições observados:

 - Grok 4.5 — 1.100 tokens de entrada, 71.500 em cache, 220 tokens de saída por requisição
- GLM-5.3/5.2/5.1 — 700 tokens de entrada, 52.000 em cache, 150 tokens de saída por requisição
+- GLM-5.2/5.1 — 700 tokens de entrada, 52.000 em cache, 150 tokens de saída por requisição
 - GPT 5.6 Luna — 1.000 tokens de entrada, 50.000 em cache, 220 tokens de saída por requisição
 - Kimi K3 — 1.050 tokens de entrada, 76.500 em cache, 300 tokens de saída por requisição
 - Kimi K2.7/K2.6 — 870 tokens de entrada, 55.000 em cache, 200 tokens de saída por requisição
@@ -143,7 +141,6 @@ As estimativas também se baseiam nos seguintes preços por 1M tokens e no uso m
 | Grok 4.5                     | $2.00   | $6.00  | $0.30            | -                | $15 |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20   | $1.20  | $0.02            | $0.25            | $15 |
 | GPT 5.6 Luna (> 272K tokens) | $0.40   | $1.80  | $0.04            | $0.50            | $15 |
-| GLM-5.3                      | $1.40   | $4.40  | $0.26            | -                | $15 |
 | GLM-5.2                      | $1.40   | $4.40  | $0.26            | -                | $60 |
 | GLM-5.1                      | $1.40   | $4.40  | $0.26            | -                | $60 |
 | Kimi K3                      | $3.00   | $15.00 | $0.30            | -                | $15 |
@@ -202,7 +199,6 @@ Você também pode acessar os modelos do Go através dos seguintes endpoints de
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -243,7 +239,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ---------------------- | ----------------- |
 | Grok 4.5          | Não usado              | 30 dias           |
 | GPT 5.6 Luna      | Não usado              | 30 dias           |
-| GLM-5.3           | Não usado              | 0 dias            |
 | GLM-5.2           | Não usado              | 0 dias            |
 | GLM-5.1           | Não usado              | 0 dias            |
 | Kimi K3           | Não usado              | 0 dias            |
@@ -90,7 +90,6 @@ Você também pode acessar nossos modelos pelos seguintes endpoints de API.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -181,7 +180,6 @@ Oferecemos um modelo pay-as-you-go. Abaixo estão os preços **por 1M tokens**.
 | 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            | -                |
-| Muse Spark 1.2                    | $1.25   | $4.25   | $0.15            | -                |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00   | $30.00  | $0.50            | $6.25            |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00  | $45.00  | $1.00            | $12.50           |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00   | $12.00  | $0.20            | $2.50            |
@@ -60,7 +60,6 @@ OpenCode Go работает так же, как и любой другой пр
 Текущий список моделей включает:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -99,7 +98,6 @@ OpenCode Go включает следующие лимиты:
 | ----------------- | ------------------- | ----------------- | ---------------- |
 | Grok 4.5          | 120                 | 300               | 600              |
 | GPT 5.6 Luna      | 2,050               | 5,100             | 10,250           |
-| GLM-5.3           | 220                 | 540               | 1,080            |
 | GLM-5.2           | 880                 | 2,150             | 4,300            |
 | GLM-5.1           | 880                 | 2,150             | 4,300            |
 | Kimi K3           | 110                 | 250               | 490              |
@@ -120,7 +118,7 @@ OpenCode Go включает следующие лимиты:
 Эти оценки основаны на наблюдаемых показателях запросов:

 - Grok 4.5 — 1,100 входных, 71,500 кешированных, 220 выходных токенов на запрос
- GLM-5.3/5.2/5.1 — 700 входных, 52,000 кешированных, 150 выходных токенов на запрос
+- GLM-5.2/5.1 — 700 входных, 52,000 кешированных, 150 выходных токенов на запрос
 - GPT 5.6 Luna — 1,000 входных, 50,000 кешированных, 220 выходных токенов на запрос
 - Kimi K3 — 1,050 входных, 76,500 кешированных, 300 выходных токенов на запрос
 - Kimi K2.7/K2.6 — 870 входных, 55,000 кешированных, 200 выходных токенов на запрос
@@ -143,7 +141,6 @@ OpenCode Go включает следующие лимиты:
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15           |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15           |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15           |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15           |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60           |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60           |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15           |
@@ -202,7 +199,6 @@ OpenCode Go включает следующие лимиты:
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -243,7 +239,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ---------------- | --------------- |
 | Grok 4.5          | Не используется  | 30 дней         |
 | GPT 5.6 Luna      | Не используется  | 30 дней         |
-| GLM-5.3           | Не используется  | 0 дней          |
 | GLM-5.2           | Не используется  | 0 дней          |
 | GLM-5.1           | Не используется  | 0 дней          |
 | Kimi K3           | Не используется  | 0 дней          |
@@ -99,7 +99,6 @@ OpenCode Zen работает как любой другой провайдер
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ https://opencode.ai/zen/v1/models
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -50,7 +50,6 @@ OpenCode Go ทำงานเหมือนกับผู้ให้บร
 รายชื่อโมเดลในปัจจุบันประกอบด้วย:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ OpenCode Go มีขีดจำกัดดังต่อไปนี้:
 | ----------------- | ---------------------- | ------------------- | ----------------- |
 | Grok 4.5          | 120                    | 300                 | 600               |
 | GPT 5.6 Luna      | 2,050                  | 5,100               | 10,250            |
-| GLM-5.3           | 220                    | 540                 | 1,080             |
 | GLM-5.2           | 880                    | 2,150               | 4,300             |
 | GLM-5.1           | 880                    | 2,150               | 4,300             |
 | Kimi K3           | 110                    | 250                 | 490               |
@@ -110,7 +108,7 @@ OpenCode Go มีขีดจำกัดดังต่อไปนี้:
 การประมาณการนี้อ้างอิงจากรูปแบบการใช้งาน request ที่สังเกตพบ:

 - Grok 4.5 — 1,100 input, 71,500 cached, 220 output tokens ต่อ request
- GLM-5.3/5.2/5.1 — 700 input, 52,000 cached, 150 output tokens ต่อ request
+- GLM-5.2/5.1 — 700 input, 52,000 cached, 150 output tokens ต่อ request
 - GPT 5.6 Luna — 1,000 input, 50,000 cached, 220 output tokens ต่อ request
 - Kimi K3 — 1,050 input, 76,500 cached, 300 output tokens ต่อ request
 - Kimi K2.7/K2.6 — 870 input, 55,000 cached, 200 output tokens ต่อ request
@@ -133,7 +131,6 @@ OpenCode Go มีขีดจำกัดดังต่อไปนี้:
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15   |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15   |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15   |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15   |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60   |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15   |
@@ -190,7 +187,6 @@ OpenCode Go มีขีดจำกัดดังต่อไปนี้:
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ----------- | ------------------ |
 | Grok 4.5          | ไม่นำไปใช้  | 30 วัน             |
 | GPT 5.6 Luna      | ไม่นำไปใช้  | 30 วัน             |
-| GLM-5.3           | ไม่นำไปใช้  | 0 วัน              |
 | GLM-5.2           | ไม่นำไปใช้  | 0 วัน              |
 | GLM-5.1           | ไม่นำไปใช้  | 0 วัน              |
 | Kimi K3           | ไม่นำไปใช้  | 0 วัน              |
@@ -92,7 +92,6 @@ OpenCode Zen ทำงานเหมือน provider อื่น ๆ ใน
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -183,7 +182,6 @@ https://opencode.ai/zen/v1/models
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -257,7 +257,7 @@ Allows the LLM to fetch and read web pages. Useful for looking up documentation
 Search the web for information.

 :::note
-This tool is only available when using the OpenCode or OpenCode Go provider, or when the `OPENCODE_ENABLE_EXA` environment variable is set to any truthy value (e.g., `true` or `1`).
+This tool is only available when using the OpenCode provider or when the `OPENCODE_ENABLE_EXA` environment variable is set to any truthy value (e.g., `true` or `1`).

 To enable when launching OpenCode:

@@ -50,7 +50,6 @@ Her çalışma alanından yalnızca bir üye OpenCode Go'ya abone olabilir.
 Mevcut model listesi şunları içerir:

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ Aşağıdaki tablo, tipik Go kullanım modellerine dayalı tahmini bir istek say
 | ----------------- | ------------------ | -------------- | ----------- |
 | Grok 4.5          | 120                | 300            | 600         |
 | GPT 5.6 Luna      | 2,050              | 5,100          | 10,250      |
-| GLM-5.3           | 220                | 540            | 1,080       |
 | GLM-5.2           | 880                | 2,150          | 4,300       |
 | GLM-5.1           | 880                | 2,150          | 4,300       |
 | Kimi K3           | 110                | 250            | 490         |
@@ -110,7 +108,7 @@ Aşağıdaki tablo, tipik Go kullanım modellerine dayalı tahmini bir istek say
 Tahminler, gözlemlenen istek modellerine dayanır:

 - Grok 4.5 — İstek başına 1.100 girdi, 71.500 önbelleğe alınmış, 220 çıktı token'ı
- GLM-5.3/5.2/5.1 — İstek başına 700 girdi, 52.000 önbelleğe alınmış, 150 çıktı token'ı
+- GLM-5.2/5.1 — İstek başına 700 girdi, 52.000 önbelleğe alınmış, 150 çıktı token'ı
 - GPT 5.6 Luna — İstek başına 1.000 girdi, 50.000 önbelleğe alınmış, 220 çıktı token'ı
 - Kimi K3 — İstek başına 1.050 girdi, 76.500 önbelleğe alınmış, 300 çıktı token'ı
 - Kimi K2.7/K2.6 — İstek başına 870 girdi, 55.000 önbelleğe alınmış, 200 çıktı token'ı
@@ -133,7 +131,6 @@ Tahminler ayrıca 1M token başına aşağıdaki fiyatlara ve her modelle birlik
 | Grok 4.5                     | $2.00  | $6.00  | $0.30       | -            | $15      |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02       | $0.25        | $15      |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04       | $0.50        | $15      |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26       | -            | $15      |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26       | -            | $60      |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26       | -            | $60      |
 | Kimi K3                      | $3.00  | $15.00 | $0.30       | -            | $15      |
@@ -190,7 +187,6 @@ Go modellerine aşağıdaki API uç noktaları aracılığıyla da erişebilirsi
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | ------------- | ------------ |
 | Grok 4.5          | Kullanılmaz   | 30 gün       |
 | GPT 5.6 Luna      | Kullanılmaz   | 30 gün       |
-| GLM-5.3           | Kullanılmaz   | 0 gün        |
 | GLM-5.2           | Kullanılmaz   | 0 gün        |
 | GLM-5.1           | Kullanılmaz   | 0 gün        |
 | Kimi K3           | Kullanılmaz   | 0 gün        |
@@ -90,7 +90,6 @@ Modellerimize aşağıdaki API uç noktaları aracılığıyla da erişebilirsin
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -181,7 +180,6 @@ Kullandıkça öde modelini destekliyoruz. Aşağıda **1M token başına** fiya
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -99,7 +99,6 @@ You can also access our models through the following API endpoints.
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -192,7 +191,6 @@ We support a pay-as-you-go model. Below are the prices **per 1M tokens**.
 | 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       | -            |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15       | -            |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50       | $6.25        |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00       | $12.50       |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20       | $2.50        |
@@ -50,7 +50,6 @@ OpenCode Go 的工作方式与 OpenCode 中的其他提供商一样。
 当前支持的模型列表包括：

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ OpenCode Go 包含以下限制：
 | ----------------- | --------------- | ---------- | ---------- |
 | Grok 4.5          | 120             | 300        | 600        |
 | GPT 5.6 Luna      | 2,050           | 5,100      | 10,250     |
-| GLM-5.3           | 220             | 540        | 1,080      |
 | GLM-5.2           | 880             | 2,150      | 4,300      |
 | GLM-5.1           | 880             | 2,150      | 4,300      |
 | Kimi K3           | 110             | 250        | 490        |
@@ -110,7 +108,7 @@ OpenCode Go 包含以下限制：
 预估值基于观察到的请求模式：

 - Grok 4.5 — 每次请求 1,100 个输入 token，71,500 个缓存 token，220 个输出 token
- GLM-5.3/5.2/5.1 — 每次请求 700 个输入 token，52,000 个缓存 token，150 个输出 token
+- GLM-5.2/5.1 — 每次请求 700 个输入 token，52,000 个缓存 token，150 个输出 token
 - GPT 5.6 Luna — 每次请求 1,000 个输入 token，50,000 个缓存 token，220 个输出 token
 - Kimi K3 — 每次请求 1,050 个输入 token，76,500 个缓存 token，300 个输出 token
 - Kimi K2.7/K2.6 — 每次请求 870 个输入 token，55,000 个缓存 token，200 个输出 token
@@ -133,7 +131,6 @@ OpenCode Go 包含以下限制：
 | Grok 4.5                     | $2.00  | $6.00  | $0.30     | -        | $15      |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02     | $0.25    | $15      |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04     | $0.50    | $15      |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26     | -        | $15      |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26     | -        | $60      |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26     | -        | $60      |
 | Kimi K3                      | $3.00  | $15.00 | $0.30     | -        | $15      |
@@ -190,7 +187,6 @@ OpenCode Go 包含以下限制：
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | -------- | -------- |
 | Grok 4.5          | 不使用   | 30 天    |
 | GPT 5.6 Luna      | 不使用   | 30 天    |
-| GLM-5.3           | 不使用   | 0 天     |
 | GLM-5.2           | 不使用   | 0 天     |
 | GLM-5.1           | 不使用   | 0 天     |
 | Kimi K3           | 不使用   | 0 天     |
@@ -90,7 +90,6 @@ OpenCode Zen 的工作方式与 OpenCode 中的任何其他提供商相同。
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -181,7 +180,6 @@ https://opencode.ai/zen/v1/models
 | 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    | -        |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15    | -        |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50    | $6.25    |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00    | $12.50   |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20    | $2.50    |
@@ -50,7 +50,6 @@ OpenCode Go 的運作方式與 OpenCode 中的任何其他供應商相同。
 目前的模型清單包括：

 - **Grok 4.5**
- **GLM-5.3**
 - **GLM-5.2**
 - **GLM-5.1**
 - **GPT 5.6 Luna**
@@ -89,7 +88,6 @@ OpenCode Go 包含以下限制：
 | ----------------- | --------------- | ---------- | ---------- |
 | Grok 4.5          | 120             | 300        | 600        |
 | GPT 5.6 Luna      | 2,050           | 5,100      | 10,250     |
-| GLM-5.3           | 220             | 540        | 1,080      |
 | GLM-5.2           | 880             | 2,150      | 4,300      |
 | GLM-5.1           | 880             | 2,150      | 4,300      |
 | Kimi K3           | 110             | 250        | 490        |
@@ -110,7 +108,7 @@ OpenCode Go 包含以下限制：
 這些預估值是基於觀察到的請求模式：

 - Grok 4.5 — 每次請求 1,100 個輸入 token、71,500 個快取 token、220 個輸出 token
- GLM-5.3/5.2/5.1 — 每次請求 700 個輸入 token、52,000 個快取 token、150 個輸出 token
+- GLM-5.2/5.1 — 每次請求 700 個輸入 token、52,000 個快取 token、150 個輸出 token
 - GPT 5.6 Luna — 每次請求 1,000 個輸入 token、50,000 個快取 token、220 個輸出 token
 - Kimi K3 — 每次請求 1,050 個輸入 token、76,500 個快取 token、300 個輸出 token
 - Kimi K2.7/K2.6 — 每次請求 870 個輸入 token、55,000 個快取 token、200 個輸出 token
@@ -133,7 +131,6 @@ OpenCode Go 包含以下限制：
 | Grok 4.5                     | $2.00  | $6.00  | $0.30     | -        | $15    |
 | GPT 5.6 Luna (≤ 272K tokens) | $0.20  | $1.20  | $0.02     | $0.25    | $15    |
 | GPT 5.6 Luna (> 272K tokens) | $0.40  | $1.80  | $0.04     | $0.50    | $15    |
-| GLM-5.3                      | $1.40  | $4.40  | $0.26     | -        | $15    |
 | GLM-5.2                      | $1.40  | $4.40  | $0.26     | -        | $60    |
 | GLM-5.1                      | $1.40  | $4.40  | $0.26     | -        | $60    |
 | Kimi K3                      | $3.00  | $15.00 | $0.30     | -        | $15    |
@@ -190,7 +187,6 @@ OpenCode Go 包含以下限制：
 | ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
 | Grok 4.5          | grok-4.5          | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
 | GPT 5.6 Luna      | gpt-5.6-luna      | `https://opencode.ai/zen/go/v1/responses`        | `@ai-sdk/openai`            |
-| GLM-5.3           | glm-5.3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.2           | glm-5.2           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | GLM-5.1           | glm-5.1           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | Kimi K3           | kimi-k3           | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
@@ -229,7 +225,6 @@ https://opencode.ai/zen/go/v1/models
 | ----------------- | -------- | -------- |
 | Grok 4.5          | 不使用   | 30 天    |
 | GPT 5.6 Luna      | 不使用   | 30 天    |
-| GLM-5.3           | 不使用   | 0 天     |
 | GLM-5.2           | 不使用   | 0 天     |
 | GLM-5.1           | 不使用   | 0 天     |
 | Kimi K3           | 不使用   | 0 天     |
@@ -94,7 +94,6 @@ OpenCode Zen 的運作方式和 OpenCode 中的其他供應商一樣。
 | 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`            |
-| Muse Spark 1.2              | muse-spark-1.2              | `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`         |
@@ -186,7 +185,6 @@ https://opencode.ai/zen/v1/models
 | 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    | -        |
-| Muse Spark 1.2                    | $1.25  | $4.25   | $0.15    | -        |
 | GPT 5.6 Sol (≤ 272K tokens)       | $5.00  | $30.00  | $0.50    | $6.25    |
 | GPT 5.6 Sol (> 272K tokens)       | $10.00 | $45.00  | $1.00    | $12.50   |
 | GPT 5.6 Terra (≤ 272K tokens)     | $2.00  | $12.00  | $0.20    | $2.50    |
Author	SHA1	Message	Date
Aiden Cline	100b2cf308	feat(opencode): route cloudflare ai gateway openai and anthropic models through native passthroughs Co-authored-by: Keefe Tang <keefe@pentagram.me>	2026-08-14 13:31:13 -05:00
Aiden Cline	387bd55583	Merge branch 'dev' into fix/ai-gateway-scope-token	2026-08-13 22:14:38 -05:00
Aiden Cline	e261ae1560	Merge branch 'dev' into fix/ai-gateway-scope-token	2026-08-11 23:59:55 -05:00
Keefe Tang	34c64fb45b	fix(provider): scope AI Gateway token to first-party Workers AI models #32052 fixed #32051 (Workers AI 401s) by passing apiKey to createUnified, but applied it to every model — so the Cloudflare API token was sent as the upstream Authorization header for third-party providers (OpenAI, Anthropic), causing them to 401 with "Invalid API Key". Scope token forwarding to be model-aware: attach the Cloudflare token only for first-party Workers AI models, whose upstream is Cloudflare itself. The Unified API addresses Workers AI both as "workers-ai/..." and as bare "@cf/..." ids, so match both; "@cf/" is Cloudflare's reserved namespace, so this never matches a third-party model. Other providers receive no upstream Authorization and fall back to the gateway's stored/BYOK keys. Applied in both the v1 provider (provider.ts) and v2 plugin (core/.../cloudflare-ai-gateway.ts) paths. Tests assert both directions, including that third-party sub-requests carry no upstream authorization header. Reapplies and extends the approach from #33407.	2026-06-24 11:52:13 +10:00