Prim - Discovery: leaderboard
Prim's primitive-discovery score: from the problem alone, the model must state its mathematical primitive, the one key idea that organizes the solution. An LLM judge grades it against a human-verified gold primitive over 182 Humanity's Last Exam mathematics problems; higher is better (Prim, 2026).
Metric: Discovery score (%): how well the model states the problem's mathematical primitive (the key organizing idea) from the problem alone, judged against the human-verified gold primitive by GPT-5.4-High on a graded rubric, over 182 problems; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked. Updated 4 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-5.4 | High | 82.42 | #138 |
| 2 | GPT-5.4 Mini | High | 59.34 | #405 |
| 3 | GPT-5.4 Nano | High | 41.76 | #651 |
| 4 | GPT-OSS-20B | High | 34.62 | #1166 |
| 5 | Qwen 3.5 27B | 28.57 | #378 | |
| 6 | Qwen 3.6 27B | 24.73 | #352 | |
| 7 | Qwen 3.5 9B | 13.19 | #829 | |
| 8 | DeepSeek-R1-Distill-Qwen-7B | 7.14 | #7912 | |
| 9 | Qwen 3.5 4B | 6.59 | #1235 | |
| 10 | DeepSeek R1 0528 Qwen3 8B | 6.59 | #2190 | |
| 11 | DeepSeek R1 Distill Qwen 32B | 6.04 | #1813 | |
| 12 | DeepSeek R1 Distill Qwen 14B | 4.4 | #2706 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=prim-discovery · How It Works · Data refreshed daily, snapshot 2026-11-20.