Released Jul 25, 2026. Everything the community has pinned to this exact release.
Codex is fast and overshoots; Claude is a pedantic reviewer. Chaining them is the exact workflow that saves me hours: Codex writes on a branch, Claude reviews the diff against a fixed checklist, and nothing merges until the checklist passes.
Same 12 tasks I run on every release. Fable took 9, Opus took 2, one tie — but the interesting part isn't the score.
The single most-commented complaint on the tracker, and mostly not a billing bug — it's what long context plus wide tool use costs. Here's what actually moved the needle for people.
Mine grew to 600 lines and the agent ignores most of it. Time to talk about what earns a slot in the most expensive real estate in your repo.
Reproduction evidence on 2.1.220
| Problem | Version | OS | Outcome | Status |
|---|---|---|---|---|
| API Error: 400 due to tool use concurrency issues. Run /rewind to recover the conversation. | 2.1.220 | Ubuntu 24.04 | Reproduced | Workaround Verified |
| API Error: 400 due to tool use concurrency issues. Run /rewind to recover the conversation. | 2.1.220 | Windows 11 | Not reproduced | Workaround Verified |
| API Error: Claude's response exceeded the 32000 output token maximum | 2.1.220 | Windows 11 | Reproduced | Workaround Verified |
| Claude Code keeps rewriting package-lock.json | 2.1.220 | macOS 15.5 | Reproduced | Workaround Verified |
| Claude Code ignores repository instructions after compaction | 2.1.220 | Windows 11 | Reproduced | Active Issue |