AA Terminal-Bench 4.0: leaderboard
Metric: Pass@1 (%). Source: artificialanalysis.ai. Saturation forecast: Around January 2027. 216 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6 Astra (xHigh) | 59.6 |
| 2 | GPT-6 Astra (Max) | 59.09 |
| 3 | Gemini 4 Argon (High) | 57.07 |
| 4 | GPT-6.1 Sol (Max) | 56.06 |
| 5 | GPT-6 Astra (High) | 54.04 |
| 6 | GPT-6.1 Sol (xHigh) | 54.04 |
| 7 | GPT-6.1 Sol (High) | 51.52 |
| 8 | GPT-6 Astra (Medium) | 49.49 |
| 9 | Claude Opus 5 (Max) | 48.99 |
| 10 | GPT-6.1 Sol (Medium) | 47.98 |
| 11 | Claude Opus 5 (xHigh) | 46.46 |
| 12 | Claude Opus 5 (High) | 45.96 |
| 13 | GPT-6 Sol (Max) | 43.94 |
| 14 | GLM-5.3 (Max) | 41.92 |
| 15 | GPT-6 Astra (Low) | 41.92 |
Interactive version: theaggregate.ai/benchmark?slug=aa-terminal-bench-4-0 · How It Works · Data refreshed daily, snapshot 2026-10-10.