Command-line Interface Tasks on Terminal-Bench 2.0
57.3Terminus2 JSON ScoreClaude-Opus-4.5
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Claude-Opus-4.5Size=?2026.02 | 57.3 | 58.4 | 53.9 | 51.7 | |
| Claude-Sonnet-4.5Size=?2026.02 | 51.7 | 51.7 | 41.6 | 37.1 | |
| Kimi-K2.5Size=1000A322026.02 | 49.4 | 38.8 | 9 | 27.5 | |
| DeepSeek-V3.2Size=671A372026.02 | 39.3 | 34.8 | — | — | |
| GLM-4.7Size=358A322026.02 | 37.1 | 44.9 | — | 31.5 | |
| Qwen3-Coder-NextSize=80A32026.02 | 36.2 | 34.2 | 30.9 | 25.8 | |
| MiniMax-M2.1Size=230A102026.02 | 32.6 | — | 42.7 | 39.3 |