Terminal-Bench 2.0 execution on Terminal-Bench 2.0 (dev)
75Pass RateClaude Code
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Claude Code2026.06 | 75 | 16.67 | |
| Arbor2026.06 | 72.22 | 13.89 | |
| Codex2026.06 | 63.89 | 5.56 | |
| Initial2026.06 | 58.33 | — |
| Method | Links | ||
|---|---|---|---|
| Claude Code2026.06 | 75 | 16.67 | |
| Arbor2026.06 | 72.22 | 13.89 | |
| Codex2026.06 | 63.89 | 5.56 | |
| Initial2026.06 | 58.33 | — |