AdvancedMathBench - ProverBench Undergraduate: leaderboard
Metric: Proof acceptance rate (%) on the 200 undergraduate problems (UG split): temperature 1.0, 64k maximum output tokens, highest available reasoning effort; each proof is accepted only if all 8 parallel passes of the authors' expert-trained automatic verifier accept it (pessimistic verification); higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 (xHigh) | 64.5 |
| 2 | Claude Opus 4.8 (Max) | 59 |
| 3 | DeepSeek V4 Pro (Max) | 54 |
| 4 | GPT-5.5 (High) | 53.3 |
| 5 | GPT-5.2 (xHigh) | 53 |
| 6 | Gemini 3.1 Pro (Preview) (High) | 46.5 |
| 7 | GLM-5.2 (Max) | 44.5 |
| 8 | Qwen 3.5 397B A17B | 40 |
| 9 | GPT-OSS-120B (High) | 20.5 |
Interactive version: theaggregate.ai/benchmark?slug=advancedmathbench-proverbench-undergraduate · How It Works · Data refreshed daily, snapshot 2026-09-29.