MathArena - ARXIVLEAN September: leaderboard
Accuracy on proving or disproving 53 September 2026 arXiv conjectures in Lean4.34.1. Submissions are checked against immutable reference statements with the strict comparator. The exact53 informal problems and formal statements are released by MathArena. Published percentage preserves the source model and effort settings; separate from informal final-answer and false-statement behavior cohorts. Source: https://matharena.ai/?comp=arxivlean--september
Metric: Accuracy (%). Source: matharena.ai. Saturation forecast: Around December 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (High) | 47.17 | #3 (Claude Opus 5.5) |
| 2 | GPT-6 Astra (Max) | 37.74 | #5 (GPT-6 Astra) |
| 3 | GPT-6.1 Sol (Max) | 35.85 | #8 (GPT-6.1 Sol) |
| 4 | DeepSeek V4.1 Flash (Max) | 5.66 | #35 (DeepSeek V4.1 Flash) |
| 5 | Qwen3.8 2.4T A95B | 1.89 | #55 |
No result here: #7 Gemini 4 Argon, #9 Claude Fable 5.1, #10 Claude Sonnet 5.5.
Interactive version: theaggregate.ai/benchmark?slug=matharena-arxivlean-september · How It Works · Data refreshed daily, snapshot 2026-10-17.