评测来源/Terminal-Bench
观察中编程
Terminal-Bench
Terminal-Bench · Terminal-Bench 的公开模型评测结果。
本来源的模型成绩
| GPT-5.6 Sol | 88.8 | 88.8 | 2026-07-09 |
| Claude Fable 5 | 88 | 88.0 | 2026-06-09 |
| GPT-5.6 Terra | 87.4 | 87.4 | 2026-07-09 |
| GPT-5.6 Luna | 84.7 | 84.7 | 2026-07-09 |
| GPT-5.5 | 84.1 | 84.1 | 2026-08-05 |
| Grok 4.5 | 83.3 | 83.3 | 2026-07-08 |
| GPT-5.5 | 82.7 | 82.7 | 2026-04-23 |
| GLM-5.2 | 82.7 | 82.7 | 2026-06-16 |
| Claude Sonnet 5 | 80.4 | 80.4 | 2026-06-30 |
| Muse Spark 1.1 | 80 | 80.0 | 2026-07-09 |
| GPT-5.5 | 78.2 | 78.2 | 2026-05-28 |
| GPT-5.5 | 75.8 | 75.8 | 2026-08-05 |
| GPT-5.4 | 75.1 | 75.1 | 2026-04-23 |
| Claude Opus 4.8 | 74.6 | 74.6 | 2026-05-28 |
| Claude Opus 4.7 | 73.8 | 73.8 | 2026-08-05 |
| GPT-5.5 | 73.4 | 73.4 | 2026-08-05 |
| Claude Opus 4.7 | 71.4 | 71.4 | 2026-08-05 |
| LongCat-2.0 | 70.8 | 70.8 | 2026-06-30 |
| Claude Opus 4.7 | 70.6 | 70.6 | 2026-08-05 |
| Gemini 3.1 Pro Preview | 70.3 | 70.3 | 2026-05-28 |
| Claude Opus 4.6 | 70.2 | 70.2 | 2026-08-05 |
| GPT-5.4 | 69.8 | 69.8 | 2026-08-05 |
| Qwen3.7 Max | 69.7 | 69.7 | 2026-05-19 |
| Gemini 3.1 Pro Preview | 68.3 | 68.3 | 2026-08-05 |
| Claude Sonnet 4.6 | 67 | 67.0 | 2026-06-30 |
| Claude Opus 4.7 | 66.1 | 66.1 | 2026-05-28 |
| MiniMax-M3 | 66 | 66.0 | 2026-06-01 |
| GLM-5.1 | 65.1 | 65.1 | 2026-08-05 |
| GPT-5.4 | 64.7 | 64.7 | 2026-08-05 |
| DeepSeek V4 Pro | 64.7 | 64.7 | 2026-08-05 |
| Kimi K2.6 | 64.3 | 64.3 | 2026-08-05 |
| Claude Sonnet 4.6 | 63.1 | 63.1 | 2026-08-05 |
| Step 3.7 Flash | 59.6 | 59.6 | 2026-05-29 |
| Nemotron 3 Ultra 550B A55B | 56.4 | 56.4 | 2026-06-04 |
| MAI-Code-1-Flash | 54.8 | 54.8 | 2026-08-05 |
| MiniMax-M2.7 | 51.1 | 51.1 | 2026-06-01 |
| Laguna XS 2.1 | 37.5 | 37.5 | 2026-07-02 |