Benchmarks / Math
ArXivMath 08/2026 (MathArena)
Research-level math problems with a checkable final answer, drawn from arXiv papers submitted in August 2026; each edition is restricted to papers published within the last month to reduce training-data contamination. Maintained by MathArena. Score: Accuracy (average performance on the competition) with a 95% confidence interval (normal approximation).
ArXivMath 08/2026 (MathArena)
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | GPT-6.1 Sol | max | 96.5% |
| 2 | GPT-6 Sol | max | 93.0% |
| 3 | GPT-6 Astra | max | 88.6% |
| 4 | Fable 5.1 | high | 87.7% |
| 5 | Opus 5.5 | high | 86.0% |
| 6 | Qwen 3.8 Max | max | 57.9% |
| 7 | Grok 4.7 | xhigh | 46.5% |
| 8 | Muse Spark 1.3 | max | 45.6% |
| 9 | DeepSeek V4.1 Flash | max | 43.9% |
| 10 | Gemini 3.8 Flash | max | 40.4% |
| 11 | Kimi K3 | max | 38.6% |