| Benchmark | Score | Top model |
|---|---|---|
| DeepSWE v1.1% | 56.6% | DeepSeek V4.1 Flash |
| GPQA Diamond% | 92.6% | GPT 5.6 Sol |
| HLE% | 43.6% | Claude Opus 5 |
| HLE (with tools)% | 56.2% | Claude Opus 5.5 |
| JobBench% | 53.4% | Claude Opus 5 |
| NL2Repo% | 55.9% | Claude Opus 5 |
| SWE-bench Pro% | 67.7% | Qwen3.8 Max |
| Terminal Bench 2.1% | 86.6% | Claude Fable 5.1 |
Figures are taken from public model cards, technical reports and vendor announcements, and are approximate where sources disagree. Results from different benchmark versions or tool-use settings never share a column: Terminal Bench 2.1, 3.0 and DeepSWE v1.1 are distinct benchmarks, and HLE with tools is separate from HLE without tools. “Top model” is the highest scoring model with a published result on that benchmark; when several models tie, all of them are listed. Benchmarks this model has no published result for are left out; the line above shows how many of the benchmarks are covered.