LLM Stats (MathVista): leaderboard
Math problems including visual contexts (geometry figures, function plots, statistical charts). LLM Stats reports text-only subset scores for text-only models.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 39 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Seed 2.1 Pro | 90.7 | #67 |
| 2 | Seed 2.1 Turbo | 90.5 | #86 |
| 3 | Seed 1.8 | 87.7 | #139 |
| 4 | O3 | 86.8 | #124 |
| 5 | O4 Mini | 84.3 | #180 |
| 6 | Step3 VL 10B | 84 | #485 |
| 7 | Command A+ | 80.6 | #413 |
| 8 | Llama 4 Maverick | 73.7 | #464 |
| 9 | GPT-4.1 Mini | 73.1 | #355 |
| 10 | GPT-4.5 | 72.3 | #274 |
| 11 | GPT-4.1 | 72.2 | #248 |
| 12 | O1 | 71.8 | #183 |
| 13 | QVQ-72B-Preview | 71.4 | #440 |
| 14 | Llama 4 Scout | 70.7 | #656 |
| 15 | Grok 2 | 69 | #452 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-mathvista · How It Works · Data refreshed daily, snapshot 2026-10-20.