fix(onboard): increase verification timeout and reduce max_tokens for custom provider probes (#27380)

* fix(onboard): increase verification timeout and reduce max_tokens for custom provider probes

The onboard wizard sends a chat-completion request to verify custom
providers.  With max_tokens: 1024 and a 10 s timeout, large local
models (e.g. Qwen3.5-27B on llama.cpp) routinely time out because
the server needs to load the model and generate up to 1024 tokens
before responding.

Changes:
- Raise VERIFY_TIMEOUT_MS from 10 s to 30 s
- Lower max_tokens from 1024 to 1 (verification only needs a single
  token to confirm the API is reachable and the model ID is valid)
- Add explicit stream: false to both OpenAI and Anthropic probes

Closes #27346

Made-with: Cursor

* Changelog: note custom-provider onboarding verification fix

---------

Co-authored-by: Philipp Spiess <hello@philippspiess.com>

src/commands/onboard-custom.test.ts

@@ -128,7 +128,7 @@ describe("promptCustomApiConfig", () => {
 
     const firstCall = fetchMock.mock.calls[0]?.[1] as { body?: string } | undefined;
     expect(firstCall?.body).toBeDefined();
-    expect(JSON.parse(firstCall?.body ?? "{}")).toMatchObject({ max_tokens: 1024 });
+    expect(JSON.parse(firstCall?.body ?? "{}")).toMatchObject({ max_tokens: 1 });
   });
 
   it("uses expanded max_tokens for anthropic verification probes", async () => {
@@ -143,7 +143,7 @@ describe("promptCustomApiConfig", () => {
     expect(fetchMock).toHaveBeenCalledTimes(2);
     const secondCall = fetchMock.mock.calls[1]?.[1] as { body?: string } | undefined;
     expect(secondCall?.body).toBeDefined();
-    expect(JSON.parse(secondCall?.body ?? "{}")).toMatchObject({ max_tokens: 1024 });
+    expect(JSON.parse(secondCall?.body ?? "{}")).toMatchObject({ max_tokens: 1 });
   });
 
   it("re-prompts base url when unknown detection fails", async () => {
@@ -220,7 +220,7 @@ describe("promptCustomApiConfig", () => {
 
     const promise = runPromptCustomApi(prompter);
 
-    await vi.advanceTimersByTimeAsync(10000);
+    await vi.advanceTimersByTimeAsync(30_000);
     await promise;
 
     expect(prompter.text).toHaveBeenCalledTimes(6);

src/commands/onboard-custom.ts

@@ -18,7 +18,7 @@ import type { SecretInputMode } from "./onboard-types.js";
 const DEFAULT_OLLAMA_BASE_URL = "http://127.0.0.1:11434/v1";
 const DEFAULT_CONTEXT_WINDOW = 4096;
 const DEFAULT_MAX_TOKENS = 4096;
-const VERIFY_TIMEOUT_MS = 10000;
+const VERIFY_TIMEOUT_MS = 30_000;
 
 /**
  * Detects if a URL is from Azure AI Foundry or Azure OpenAI.
@@ -317,7 +317,8 @@ async function requestOpenAiVerification(params: {
     body: {
       model: params.modelId,
       messages: [{ role: "user", content: "Hi" }],
-      max_tokens: 1024,
+      max_tokens: 1,
+      stream: false,
     },
   });
 }
@@ -343,8 +344,9 @@ async function requestAnthropicVerification(params: {
     headers: buildAnthropicHeaders(params.apiKey),
     body: {
       model: params.modelId,
-      max_tokens: 1024,
+      max_tokens: 1,
       messages: [{ role: "user", content: "Hi" }],
+      stream: false,
     },
   });
 }