LLM Stats (MathVista-Mini): leaderboard
1,000-question subset of MathVista covering math problems with visual inputs including geometry, charts, function plots, and scientific figures.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.5 | 90.1 | #144 |
| 2 | Qwen 3.5 27B | 87.8 | #187 |
| 3 | Qwen 3.5 122B A10B | 87.4 | #173 |
| 4 | Qwen 3.6 27B | 87.4 | #170 |
| 5 | Qwen 3.6 35B A3B | 86.4 | #226 |
| 6 | Qwen 3.5 35B A3B | 86.2 | #256 |
| 7 | Qwen 3 VL 32B (Thinking) | 85.9 | #294 (Qwen 3 VL 32B) |
| 8 | Qwen 3 VL 235B A22B (Thinking) | 85.8 | #238 (Qwen 3 VL 235B A22B) |
| 9 | Seed 2.0 Mini | 85.5 | #276 |
| 10 | EXAONE 4.5 33B | 85 | #396 |
| 11 | Qwen 3 VL 235B A22B Instruct | 84.9 | #273 |
| 12 | Qwen 3 VL 32B Instruct | 83.8 | #283 |
| 13 | Qwen 3 VL 30B A3B (Thinking) | 81.9 | #351 (Qwen 3 VL 30B A3B) |
| 14 | Qwen 3 VL 8B (Thinking) | 81.4 | |
| 15 | Qwen 3 VL 30B A3B Instruct | 80.1 | #384 |
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-mini · How It Works · Data refreshed daily, snapshot 2026-10-20.