Vals AI Terminal-Bench 2.1: leaderboard
Updated Terminal-Bench 2.1 evaluation from Vals AI, measuring agentic command-line task completion in sandboxed software and systems environments.
Metric: Accuracy (%). Source: www.vals.ai. Saturation forecast: Around January 2027. 76 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 | 87.64 | #3 |
| 2 | GPT-6 Astra | 87.27 | #6 |
| 3 | GPT-5.6 Sol | 85.77 | #19 |
| 4 | Claude Fable 5.1 | 85.02 | #9 |
| 5 | Claude Opus 5 | 84.64 | #12 |
| 6 | GPT-6 Sol | 83.15 | #14 |
| 7 | Claude Sonnet 5.5 | 83.15 | #11 |
| 8 | Gemini 3.8 Flash | 81.27 | #21 |
| 9 | Kimi K3 | 80.9 | #30 |
| 10 | Claude Fable 5 | 80.52 | #16 |
| 11 | GPT-5.6 Luna | 79.03 | #68 |
| 12 | Muse Spark 1.3 (Max) | 79.03 | #24 (Muse Spark 1.3) |
| 13 | Grok 4.6 | 78.28 | #26 |
| 14 | GPT-5.6 Terra | 77.53 | #41 |
| 15 | Gemini 3.7 Flash | 77.53 | #23 |
No result here: #7 GPT-6.1 Sol, #8 Gemini 4 Argon, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=vals-ai-terminal-bench-2-1 · How It Works · Data refreshed daily, snapshot 2026-10-20.