Terminal-Bench 4.0 (Codex): leaderboard
Metric: Resolution Rate (%). Source: www.tbench.ai. Saturation forecast: Around July 2027. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6 Astra (Max) | 58.18 |
| 2 | GPT-6.1 Sol (Max) | 58.18 |
| 3 | GPT-6 Astra (High) | 57.88 |
| 4 | GPT-6 Astra (xHigh) | 57.88 |
| 5 | GPT-6 Astra (Medium) | 54.24 |
| 6 | GPT-6 Astra (Low) | 50.61 |
| 7 | GPT-6 Sol (Max) | 49.39 |
| 8 | GPT-5.6 Sol (Max) | 37.27 |
| 9 | GPT-5.6 Terra (Max) | 21.52 |
| 10 | GPT-5.6 Luna (Max) | 17.27 |
| 11 | GPT-6 Luna (Max) | 16.36 |
Interactive version: theaggregate.ai/benchmark?slug=terminal-bench-4-0-codex · How It Works · Data refreshed daily, snapshot 2026-10-09.