LiveMathematicianBench — leaderboard
Continuously updated benchmark testing LLMs on understanding cutting-edge mathematical theorems from newly published arXiv preprints. Multiple-choice questions constructed from theorem statements and proof sketches.
Metric: Accuracy (%). Source: livemathematicianbench.github.io. Status: saturation imminent. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 58.6 |
| 2 | Gemini 3.1 Pro (Preview) | 54.3 |
| 3 | Kimi K2.5 | 48.3 |
| 4 | Qwen 3.5 397B A17B | 47.8 |
| 5 | GPT-OSS-120B | 34.8 |
| 6 | MiniMax-M2.5 | 31 |
| 7 | Grok 4.1 Fast (Reasoning) | 30.8 |
Interactive version: theaggregate.ai/benchmark?slug=livemathematicianbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.