Benchmarks · math
MathArena
Uses MathArena's website-wide IRT expected-performance estimate. Score: website-wide IRT expected performance. Collected by Jake Boggs' benchmark dashboard.
88 models · 95 results · data as of 2026-11-06 · matharena.ai ↗
Leaderboard
GPT-6.1 Sol max
All results
95 rows| # | Model | Effort | MathArena |
|---|---|---|---|
| 1 | GPT-6.1 Sol | max | 84.0% |
| 2 | GPT-6 Astra | max | 81.8% |
| 3 | GPT-6 Sol | max | 79.9% |
| 4 | Opus 5.5 | max | 79.0% |
| 5 | Fable 5.1 | max | 77.9% |
| 6 | GPT-6 Astra | low | 77.1% |
| 7 | Gemini 4 Argon | max | 70.0% |
| 8 | Opus 5 | max | 69.5% |
| 9 | GPT-5.6 Sol | max | 61.3% |
| 10 | GPT-5.5 | xhigh | 59.8% |
| 11 | Kimi K3.1 | max | 59.6% |
| 12 | Muse Wonder (Max) | max | 56.4% |
| 13 | GPT-5.4 Pro | xhigh | 51.9% |
| 14 | DeepSeek V4.1 Flash 1017 | max | 51.6% |
| 15 | Qwen 3.8 Max | max | 51.0% |
| 16 | Fable 5 | max | 50.9% |
| 17 | Kimi K3 | max | 47.5% |
| 18 | Muse Spark 1.3 | max | 42.5% |
| 19 | Muse Spark 1.1 | max | 42.4% |
| 20 | DeepSeek V4.1 Flash | max | 42.3% |
| 21 | Grok 4.7 | xhigh | 40.7% |
| 22 | Opus 4.8 | max | 39.9% |
| 23 | GPT-5.4 | xhigh | 37.4% |
| 24 | GPT-5.2 | xhigh | 34.1% |
| 25 | Grok 4.5 | max | 34.1% |
| 26 | Gemini 3.8 Flash | max | 33.6% |
| 27 | Muse Spark Open | xhigh | 33.1% |
| 28 | Qwen 3.8 | xhigh | 32.8% |
| 29 | Gemini 3.1 Pro | high | 30.2% |
| 30 | Mistral Large 4 | max | 26.2% |