LLM Stats (CritPT): leaderboard
CritPt (Complex Research using Integrated Thinking - Physics Test): unpublished research-level physics challenges written by working physicists across modern research areas, testing open-ended reasoning well beyond textbook problems.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around September 2028. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Hy4 preview | 16.9 | #38 |
| 2 | GLM-5.2 | 16.7 | #73 |
| 3 | Qwen 3.7 Max | 11.4 | #75 |
| 4 | Inkling Small | 8.3 | #138 |
| 5 | Qwen 3.7 Plus | 6 | #92 |
| 6 | Nemotron 3 Ultra | 3.1 | #235 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-critpt · How It Works · Data refreshed daily, snapshot 2026-10-20.