fix(codex): handle fast tier and capacity SSE (#2452)

- map service_tier=fast to upstream priority; drop unsupported tiers
- normalize reasoning effort max to xhigh (codex-only)
- convert 200-SSE model-capacity errors into 503 so account fallback rotates
- keep normal SSE output intact after peeking

Co-authored-by: Cursor <cursoragent@cursor.com>

tests/unit/codex-fast-capacity.test.js

@@ -0,0 +1,71 @@
+import { describe, expect, it } from "vitest";
+import { CodexExecutor } from "../../open-sse/executors/codex.js";
+
+function streamFromText(text) {
+  const encoder = new TextEncoder();
+  return new ReadableStream({
+    start(controller) {
+      controller.enqueue(encoder.encode(text));
+      controller.close();
+    },
+  });
+}
+
+describe("Codex fast tier and capacity handling", () => {
+  it("maps Codex fast tier to priority and max reasoning to xhigh", () => {
+    const executor = new CodexExecutor();
+    const body = executor.transformRequest("gpt-5.5", {
+      model: "gpt-5.5",
+      input: "hi",
+      reasoning_effort: "max",
+      service_tier: "fast",
+    }, true, {});
+
+    expect(body.service_tier).toBe("priority");
+    expect(body.reasoning.effort).toBe("xhigh");
+  });
+
+  it("uses ChatGPT workspace header fallback", () => {
+    const executor = new CodexExecutor();
+    const headers = executor.buildHeaders({
+      accessToken: "token",
+      connectionId: "conn_1",
+      providerSpecificData: { chatgptAccountId: "acct_1" },
+    });
+
+    expect(headers["ChatGPT-Account-ID"]).toBe("acct_1");
+  });
+
+  it("classifies 200-SSE model capacity as account fallback", async () => {
+    const executor = new CodexExecutor();
+    const response = new Response(streamFromText([
+      "event: error",
+      'data: {"error":{"message":"Selected model is at capacity. Please try a different model."}}',
+      "",
+    ].join("\n")), {
+      status: 200,
+      headers: { "Content-Type": "text/event-stream" },
+    });
+
+    const peek = await executor._peekSseTransientError(response);
+    expect(peek.accountFallback).toBe(true);
+    expect(peek.message).toBe("Selected model is at capacity. Please try a different model.");
+  });
+
+  it("reassembles normal SSE after peeking", async () => {
+    const executor = new CodexExecutor();
+    const text = [
+      "event: response.output_text.delta",
+      'data: {"type":"response.output_text.delta","delta":"OK"}',
+      "",
+    ].join("\n");
+    const response = new Response(streamFromText(text), {
+      status: 200,
+      headers: { "Content-Type": "text/event-stream" },
+    });
+
+    const peek = await executor._peekSseTransientError(response);
+    expect(peek.matched).toBeNull();
+    await expect(new Response(peek.replacementBody).text()).resolves.toBe(text);
+  });
+});