The original Terminal-Bench — tests LLMs on terminal/CLI tasks including shell scripting, file operations, and system administration commands.
This definition has not completed editorial and provenance review. Verify its metric, version, source, and evaluation configuration upstream before using it for a model-selection decision.
Established baseline for CLI competency. Simpler tasks than v2.0 provide better coverage of basic shell skills.
Superseded by Terminal-Bench 2.0 which has harder, more realistic tasks. Lower difficulty ceiling.
Higher is better. Task completion rate (0–100%).
Superseded by Terminal-Bench 2.0 which has harder, more realistic tasks.
| # | Model | Score |
|---|---|---|
| 1 | 87.8 | |
| 2 | 78.9 | |
| 3 | 74.6 | |
| 4 | 63.2 | |
| 5 | 61.7 | |
| 6 | 58.5 | |
| 7 | 57.9 | |
| 8 | 55.9 | |
| 9 | 55.6 | |
| 10 | 53.9 | |
| 11 | 53.8 | |
| 12 | 53.7 | |
| 13 | 53.4 | |
| 14 | 52.5 | |
| 15 | 52.5 | |
| 16 | 52.3 | |
| 17 | 51.0 | |
| 18 | 51.0 | |
| 19 | 49.2 | |
| 20 | 48.4 | |
| 21 | 48.0 | |
| 22 | 47.8 | |
| 23 | 47.5 | |
| 24 | 47.1 | |
| 25 | 47.0 | |
| 26 | 46.6 | |
| 27 | 46.4 | |
| 28 | 46.4 | |
| 29 | 46.3 | |
| 30 | 45.8 | |
| 31 | 45.7 | |
| 32 | 45.4 | |
| 33 | 45.2 | |
| 34 | 45.2 | |
| 35 | 45.2 | |
| 36 | 45.0 | |
| 37 | 44.9 | |
| 38 | 44.5 | |
| 39 | 44.5 | |
| 40 | 44.4 | |
| 41 | 44.1 | |
| 42 | 44.0 | |
| 43 | 44.0 | |
| 44 | 43.9 | |
| 45 | 43.8 | |
| 46 | 43.8 | |
| 47 | 43.7 | |
| 48 | 43.5 | |
| 49 | 43.0 | |
| 50 | 42.6 |