Terminal-Bench 2.0 on qwen3-coder-480b-a35b-instruct-fp8
89 tasks, 5 attempts, open network. Updated 25 Oct 2026.
Ranking
Pass rate by harness, higher is better
All harnesses
| Harness | pass^5 | Runs | Last published |
|---|---|---|---|
| aone-agent1.18.72-pre | 11.2% | 13 infra | 25 Oct 2026run #93 |