Benchmarks · reasoning
CritPt v2
Uses Artificial Analysis accuracy on the revised research-level physics benchmark, evaluated with code execution. Score: accuracy. Collected by Jake Boggs' benchmark dashboard.
48 models · 75 results · data as of 2026-11-06 · artificialanalysis.ai ↗
Leaderboard
GPT-6 Astra medium
All results
75 rows| # | Model | Effort | CritPt v2 |
|---|---|---|---|
| 1 | GPT-6 Astra | medium | 84.0% |
| 2 | GPT-6 Astra | high | 83.7% |
| 3 | GPT-6 Astra | xhigh | 83.4% |
| 4 | GPT-6 Astra | max | 83.1% |
| 5 | GPT-6.1 Sol | max | 82.6% |
| 6 | GPT-6.1 Sol | xhigh | 82.6% |
| 7 | Gemini 4 Carbon | high | 82.3% |
| 8 | Gemini 4 Carbon | medium | 82.3% |
| 9 | GPT-6.1 Sol | high | 82.0% |
| 10 | Opus 5.5 | max | 81.7% |
| 11 | GPT-6.1 Terra | max | 81.4% |
| 12 | GPT-6 Sol | max | 80.9% |
| 13 | GPT-6.1 Sol | medium | 80.9% |
| 14 | Gemini 4 Carbon | low | 79.7% |
| 15 | GPT-6 Astra | low | 78.9% |
| 16 | GPT-6.1 Sol | low | 78.6% |
| 17 | Fable 5.1 | max | 77.7% |
| 18 | GPT-5.6 Sol | max | 77.1% |
| 19 | GPT-6.1 Terra | xhigh | 77.1% |
| 20 | GPT-6.1 Terra | high | 76.9% |
| 21 | Sonnet 5.5 | max | 75.7% |
| 22 | Muse Wonder Full | max | 75.1% |
| 23 | GPT-6.1 Terra | medium | 74.6% |
| 24 | Gemini 4 Argon | high | 72.9% |
| 25 | Opus 5 | max | 71.4% |
| 26 | Muse Wonder (Max) | max | 71.1% |
| 27 | Kimi K3.1 | max | 70.6% |
| 28 | GPT-5.6 Terra | max | 69.7% |
| 29 | GPT-6.1 Terra | low | 68.6% |
| 30 | Muse Spark 1.3 | max | 68.0% |