fix(codex): remove invalid xhigh effort option for gpt-5.5

[vdekrijger]

Problem

xhigh ("Extra High") is not a valid reasoning_effort value for the GPT 5.5 family in OpenAI's Codex API, but our app was advertising it as a selectable option — so picking it would fail.

Changes

Removed the xhigh reasoning-effort tier from the Codex adapter; it now caps at high for all models. Claude models (Opus/Fable) still support xhigh, so that path is unchanged.

How did you test this?

• pnpm --filter @posthog/agent test for the codex models, reasoning-effort, and agent suites
• pnpm --filter @posthog/workspace-server test for the agent service suite
• pnpm --filter @posthog/agent typecheck

Automatic notifications

• Publish to changelog?
• Alert Sales and Marketing teams?

---

Created with PostHog Code from a Slack thread

[github-actions]

React Doctor found no issues in the changed files. 🎉

_{Reviewed by React Doctor for commit 898d170.}

[greptile-apps]

Comments Outside Diff (1)

1. packages/agent/src/adapters/reasoning-effort.test.ts, line 5-13 (link)

   The three assertions are testing the same property (xhigh rejection) across different model IDs. The codebase preference is for parameterised tests — it.each would also make it easier to add new model IDs in the future, consistent with the pattern already used in models.test.ts.

   Context Used: Do not attempt to comment on incorrect alphabetica... (source)

   Prompt To Fix With AI

   This is a comment left during a code review.
   Path: packages/agent/src/adapters/reasoning-effort.test.ts
   Line: 5-13
   
   Comment:
   The three assertions are testing the same property (`xhigh` rejection) across different model IDs. The codebase preference is for parameterised tests — `it.each` would also make it easier to add new model IDs in the future, consistent with the pattern already used in `models.test.ts`.
   
   
   
   **Context Used:** Do not attempt to comment on incorrect alphabetica... ([source](https://app.greptile.com/review/custom-context?memory=instruction-0))
   
   How can I resolve this? If you propose a fix, please make it concise.

   Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

Prompt To Fix All With AI

Fix the following 1 code review issue. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 1
packages/agent/src/adapters/reasoning-effort.test.ts:5-13
The three assertions are testing the same property (`xhigh` rejection) across different model IDs. The codebase preference is for parameterised tests — `it.each` would also make it easier to add new model IDs in the future, consistent with the pattern already used in `models.test.ts`.

```suggestion
  it.each(["gpt-5.5", "gpt-5.5-codex", "gpt-5.3-codex"])(
    "rejects xhigh for codex models, including the gpt-5.5 family (%s)",
    (modelId) => {
      expect(isSupportedReasoningEffort("codex", modelId, "xhigh")).toBe(false);
    },
  );
```

_{Reviews (1): Last reviewed commit: "fix(codex): remove invalid xhigh effort ..." | Re-trigger Greptile}