Benchmarks · reasoning
CritPt v2
CritPt v2: the authors' revision of CritPt's 70 research-level physics challenges, solved with a code-execution tool in an offline Python sandbox; accuracy.
Category
28 models · 71 results · data as of 2026-10-29 · artificialanalysis.ai ↗
Leaderboard
GPT-6 Astra high
More from this benchmark
4 published measuresScore vs any published measure4 measures: sub-scores, splits, counts
Breakdown
darker is better within each column · click a column to sort| Model | ||||
|---|---|---|---|---|
| GPT-6 Astra high | 1,872,561 | 238,790 | 403,872 | 1,324,272 |
| GPT-6.1 Sol max | 2,576,235 | 244,360 | 838,273 | 1,737,568 |
| GPT-6.1 Terra max | 13,272,461 | 304,897 | 1,965,803 | 11,219,783 |
| Opus 5.5 xhigh, fallback | 10,948,930 | 401,053 | 1,403,006 | 9,082,614 |
| Gemini 4 Carbon high | 92,232,418 | 1,716,018 | 5,465,333 | 84,159,671 |
| Fable 5.1 high, fallback | 10,125,511 | 378,189 | 1,512,055 | 8,201,499 |
| GPT-5.6 Sol max | 12,378,515 | 301,323 | 1,513,344 | 10,756,271 |
| Sonnet 5.5 max, fallback | 69,320,837 | 765,698 | 6,808,534 | 61,792,719 |
| Gemini 4 Argon high | 52,153,368 | 918,817 | 2,986,496 | 47,781,045 |
| Muse Spark 1.3 max | 58,166,511 | 982,147 | 3,999,604 | 53,027,490 |
| Haiku 5.5 xhigh | 32,524,438 | 542,641 | 4,403,413 | 27,772,163 |
| GPT-6 Luna max | 18,055,710 | 256,864 | 2,605,477 | 16,252,349 |
| Gemini 3.8 Flash medium | 69,985,999 | 1,122,166 | 2,008,531 | 66,490,968 |
| Kimi K3 max | 31,636,958 | 428,578 | 2,200,053 | 28,837,868 |
| Grok 4.7 low | 44,607,223 | 693,605 | 2,090,883 | 41,426,082 |
| GLM 5.3 max | 139,272,171 | 916,572 | 8,952,093 | 129,182,373 |
| MiMo V2.6 Pro | 86,649,636 | 725,647 | 6,060,033 | 79,479,748 |
| MiMo V2.6 Flash | 109,666,133 | 785,115 | 7,614,710 | 100,662,310 |
| DeepSeek V4.1 Flash 1017 max | 125,075,164 | 814,225 | 7,226,416 | 116,718,406 |
| GLM 5.3 Flash max | 239,807,541 | 1,050,201 | 20,034,414 | 219,874,162 |
| MiniMax M3 | 137,062,019 | 890,449 | 7,210,133 | 128,568,280 |
| Inkling Small | 48,857,553 | 741,403 | 3,901,659 | 44,391,611 |
| Inkling xhigh | 68,466,802 | 800,065 | 4,969,454 | 62,822,918 |
| GPT-5 high | 9,036,297 | 414,438 | 1,773,888 | 7,653,681 |
| Qwen 3.8 27B medium | 152,199,849 | 1,385,504 | 7,304,410 | 142,867,123 |
All results
71 rows| # | Model | Effort | CritPt v2 | Harness |
|---|---|---|---|---|
| 1 | GPT-6 Astra | high | 83.7% | Stirrup |
| 2 | GPT-6 Astra | medium | 83.4% | Stirrup |
| 3 | GPT-6 Astra | max | 83.1% | Stirrup |
| 4 | GPT-6 Astra | xhigh | 83.1% | Stirrup |
| 5 | GPT-6.1 Sol | max | 82.0% | Stirrup |
| 6 | GPT-6.1 Sol | xhigh | 81.7% | Stirrup |
| 7 | GPT-6.1 Sol | high | 81.4% | Stirrup |
| 8 | GPT-6.1 Sol | medium | 79.7% | Stirrup |
| 9 | GPT-6.1 Terra | max | 79.4% | Stirrup |
| 10 | Opus 5.5 | xhigh | 78.9% | Stirrup |
| 11 | Gemini 4 Carbon | high | 78.9% | Stirrup |
| 12 | Opus 5.5 | max | 77.7% | Stirrup |
| 13 | GPT-6 Astra | low | 77.1% | Stirrup |
| 14 | Fable 5.1 | high | 76.9% | Stirrup |
| 15 | GPT-6.1 Sol | low | 76.9% | Stirrup |
| 16 | GPT-6.1 Terra | high | 76.3% | Stirrup |
| 17 | GPT-6.1 Terra | xhigh | 75.4% | Stirrup |
| 18 | Opus 5.5 | high | 75.1% | Stirrup |
| 19 | GPT-5.6 Sol | max | 74.9% | Stirrup |
| 20 | Fable 5.1 | xhigh | 74.6% | Stirrup |
| 21 | GPT-6.1 Terra | medium | 74.3% | Stirrup |
| 22 | Sonnet 5.5 | max | 74.0% | Stirrup |
| 23 | Sonnet 5.5 | xhigh | 73.1% | Stirrup |
| 24 | Gemini 4 Argon | high | 72.9% | Stirrup |
| 25 | Fable 5.1 | medium | 72.3% | Stirrup |
| 26 | Fable 5.1 | max | 72.0% | Stirrup |
| 27 | Opus 5.5 | medium | 71.4% | Stirrup |
| 28 | Sonnet 5.5 | high | 67.7% | Stirrup |
| 29 | GPT-6.1 Terra | low | 67.4% | Stirrup |
| 30 | Muse Spark 1.3 | max | 66.9% | Stirrup |