fix(github-copilot): honor live prompt token limits (#103275)

* fix(github-copilot): honor live prompt token limits

* fix(github-copilot): align prompt token budgets

* fix(github-copilot): align prompt token budgets

---------

Co-authored-by: saju01 <saju@coderedcorp.com>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Co-authored-by: Peter Steinberger <peter@steipete.me>

extensions/github-copilot/index.ts

@@ -375,8 +375,8 @@ export default definePluginEntry({
         }
       }
       // Try to fetch the live model catalog from Copilot's /models endpoint so
-      // the runtime tracks per-account entitlements and accurate context
-      // windows (max_context_window_tokens) without manifest churn. On any
+      // the runtime tracks per-account entitlements and accurate token limits
+      // without manifest churn. On any
       // failure we return an empty model list, which lets the static manifest
       // catalog continue to be the visible fallback for users.
       let discoveredModels: Awaited<ReturnType<typeof fetchCopilotModelCatalog>> = [];

extensions/github-copilot/model-metadata.ts

@@ -47,6 +47,7 @@ const STATIC_MODEL_OVERRIDES = new Map<string, Partial<ModelDefinitionConfig>>([
       name: "GPT-5.5",
       reasoning: true,
       contextWindow: 400_000,
+      contextTokens: 272_000,
       maxTokens: 128_000,
     },
   ],

extensions/github-copilot/models.test.ts

@@ -126,6 +126,7 @@ describe("resolveCopilotForwardCompatModel", () => {
       input: ["text", "image"],
       cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
       contextWindow: 400_000,
+      contextTokens: 272_000,
       maxTokens: 128_000,
     });
   });
@@ -570,6 +571,7 @@ describe("fetchCopilotModelCatalog", () => {
       input: ["text", "image"],
       cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
       contextWindow: 400000,
+      contextTokens: 272000,
       maxTokens: 128000,
       compat: { supportedReasoningEfforts: ["low", "medium", "high"] },
     });
@@ -578,6 +580,7 @@ describe("fetchCopilotModelCatalog", () => {
     expect(codex?.input).toEqual(["text"]);
     expect(codex?.reasoning).toBe(true);
     expect(codex?.contextWindow).toBe(400000);
+    expect(codex?.contextTokens).toBeUndefined();
 
     const gemini = out.find((m) => m.id === "gemini-3.1-pro-preview");
     expect(gemini?.api).toBe("openai-completions");
@@ -664,6 +667,7 @@ describe("fetchCopilotModelCatalog", () => {
               type: "chat",
               limits: {
                 max_context_window_tokens: -1,
+                max_prompt_tokens: -1,
                 max_output_tokens: 128000.5,
               },
             },
@@ -676,6 +680,7 @@ describe("fetchCopilotModelCatalog", () => {
               type: "chat",
               limits: {
                 max_context_window_tokens: Number.POSITIVE_INFINITY,
+                max_prompt_tokens: Number.POSITIVE_INFINITY,
                 max_output_tokens: 0,
               },
             },
@@ -696,11 +701,13 @@ describe("fetchCopilotModelCatalog", () => {
       contextWindow: 128000,
       maxTokens: 8192,
     });
+    expect(out[0]).not.toHaveProperty("contextTokens");
     expect(out[1]).toMatchObject({
       id: "gpt-bad-output",
       contextWindow: 128000,
       maxTokens: 8192,
     });
+    expect(out[1]).not.toHaveProperty("contextTokens");
   });
 
   it("throws on non-2xx HTTP responses so the caller can fall back to the static catalog", async () => {

extensions/github-copilot/models.ts

@@ -80,6 +80,9 @@ export function resolveCopilotForwardCompatModel(
       input: staticOverride.input ?? ["text", "image"],
       cost: staticOverride.cost ?? { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
       contextWindow: staticOverride.contextWindow ?? DEFAULT_CONTEXT_WINDOW,
+      ...(staticOverride.contextTokens !== undefined
+        ? { contextTokens: staticOverride.contextTokens }
+        : {}),
       maxTokens: staticOverride.maxTokens ?? DEFAULT_MAX_TOKENS,
       ...(staticOverride.thinkingLevelMap
         ? { thinkingLevelMap: staticOverride.thinkingLevelMap }
@@ -218,6 +221,7 @@ function mapCopilotApiModelToDefinition(
 
   const contextWindow =
     asPositiveSafeInteger(limits?.max_context_window_tokens) ?? DEFAULT_CONTEXT_WINDOW;
+  const contextTokens = asPositiveSafeInteger(limits?.max_prompt_tokens);
   const maxTokens = asPositiveSafeInteger(limits?.max_output_tokens) ?? DEFAULT_MAX_TOKENS;
   const compat = mergeCopilotCompat(resolveCopilotModelCompat(id), supports?.reasoning_effort);
   const api = resolveCopilotApiForVendor(entry.vendor, id);
@@ -231,6 +235,7 @@ function mapCopilotApiModelToDefinition(
     input,
     cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
     contextWindow,
+    ...(contextTokens !== undefined ? { contextTokens } : {}),
     maxTokens,
     ...(thinkingLevelMap ? { thinkingLevelMap } : {}),
     ...(compat ? { compat } : {}),

extensions/github-copilot/openclaw.plugin.json

@@ -108,6 +108,7 @@
             "reasoning": true,
             "input": ["text", "image"],
             "contextWindow": 400000,
+            "contextTokens": 272000,
             "maxTokens": 128000,
             "cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }
           },