Results
Results
Every published result, one row per harness version on each board. Filter it, or download what you see as CSV.
Select two rows of one board to compare them task by task.
| Harness | Board | Pass rateHigher is better, 95% interval | Runs | Cost per task | hbench | Last finished |
|---|---|---|---|---|---|---|
| Terminal-Bench 2.0 on gpt-5-nano (5 attempts, open network) | 21.8%15.5–28.1 | 1 | — | converted | 24 Mar 2026 | |
| Terminal-Bench 2.0 on Gemini 3 Flash Preview (5 attempts, open network) | 47.4%38.7–55.7 | 1 | — | converted | 5 Mar 2026 | |
Terminus 22.0.0 | Terminal-Bench 2.0 on deepseek-v3.2 (5 attempts, open network) | 39.6%31.5–48.1 | 1 | $0.03 | converted | 8 Feb 2026 |
hookele0.0.1 | Terminal-Bench 2.0 on gpt-5.1-codex-mini (5 attempts, open network) | 61.1%51.2–70.1 | 1 | — | converted | 6 Feb 2026 |
vixv0.1.22 | Terminal-Bench 2.0 on claude-opus-4-7 (5 attempts, open network) | 89.9%84.7–93.9 | 1 | — | converted | 1 May 2026 |
Claude Code2.1.34 | Terminal-Bench 2.0 on GLM-4.7 (5 attempts, open network) | 33.3%25.2–41.6 | 1 | — | converted | 7 Feb 2026 |