Agentic terminal task resolution on Terminal-Bench 2
83.14SSA Pass@1 RateOpus 4.6
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Opus 4.6Evaluation Setting=Unconstrained2026.06 | 83.14 | — | |
| GPT-5.3Evaluation Setting=Unconstrained2026.06 | 80.67 | — | |
| GPT-5.4Evaluation Setting=Unconstrained2026.06 | 79.1 | — | |
| Gemini 3.1 ProEvaluation Setting=Unconstrained2026.06 | 76 | — | |
| GPT-5.3Evaluation Setting=Constrained2026.06 | 75.05 | 77.3 | |
| GPT-5.2Evaluation Setting=Unconstrained2026.06 | 74.6 | — | |
| Sonnet 4.6Evaluation Setting=Unconstrained2026.06 | 74.15 | — | |
| Opus 4.6Evaluation Setting=Constrained2026.06 | 73.93 | 65.4 | |
| GPT-5.2-codexEvaluation Setting=Unconstrained2026.06 | 73.7 | — | |
| GPT-5.4Evaluation Setting=Constrained2026.06 | 72.58 | 75.1 | |
| Opus 4.5Evaluation Setting=Unconstrained2026.06 | 72.35 | — | |
| Gemini 3.1 ProEvaluation Setting=Constrained2026.06 | 72.13 | 68.5 | |
| GPT-5.2-codexEvaluation Setting=Constrained2026.06 | 68.76 | 64.7 | |
| Gemini 3 FlashEvaluation Setting=Unconstrained2026.06 | 66.29 | — | |
| Opus 4.5Evaluation Setting=Constrained2026.06 | 65.39 | 59.8 | |
| Sonnet 4.6Evaluation Setting=Constrained2026.06 | 64.71 | 59.1 | |
| Gemini 3 FlashEvaluation Setting=Constrained2026.06 | 63.59 | 47.6 | |
| GPT-5.2Evaluation Setting=Constrained2026.06 | 62.69 | 62.2 | |
| Grok 4.2Evaluation Setting=Unconstrained2026.06 | 56.4 | — | |
| Qwen3.6 27BEvaluation Setting=Unconstrained2026.06 | 56 | 59.3 | |
| Sonnet 4.5Evaluation Setting=Unconstrained2026.06 | 54.83 | — | |
| Grok 4.2Evaluation Setting=Constrained2026.06 | 52.35 | — | |
| Qwen3.5 27BEvaluation Setting=Unconstrained2026.06 | 50.78 | 41.6 | |
| Sonnet 4.5Evaluation Setting=Constrained2026.06 | 49.21 | 51 | |
| Qwen3.6 35B-A3BEvaluation Setting=Unconstrained2026.06 | 48.76 | 51.5 | |
| Haiku 4.5Evaluation Setting=Unconstrained2026.06 | 43.37 | — | |
| Sonnet 4.0Evaluation Setting=Unconstrained2026.06 | 42.47 | — | |
| Sonnet 4.0Evaluation Setting=Constrained2026.06 | 41.8 | — | |
| Haiku 4.5Evaluation Setting=Constrained2026.06 | 41.57 | 41 | |
| Qwen3-Coder NextEvaluation Setting=Unconstrained2026.06 | 38.87 | 36.2 | |
| GPT-OSS 120BEvaluation Setting=Unconstrained2026.06 | 36.17 | — | |
| Qwen3-Coder 480BEvaluation Setting=Unconstrained2026.06 | 35.73 | — | |
| GPT-OSS 20BEvaluation Setting=Unconstrained2026.06 | 30.11 | — | |
| Qwen3-Coder 30BEvaluation Setting=Unconstrained2026.06 | 17.07 | — |