HELM VHELM - Math Vista - Elementary School (Free Form): leaderboard
VHELM (Stanford CRFM HELM for vision-language models, run v2.1.2) MathVista items at grade level elementary school in free form format: mathematical reasoning over plots, charts, diagrams and photos; 151 questions, exact match. Percent correct, 0-100, higher is better.
Metric: Exact Match (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 48 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 74.17 | #126 |
| 2 | O3 (2025-04-16) | 74.17 | #125 |
| 3 | O1 (2024-12-17) | 65.56 | #157 |
| 4 | O4 Mini (2025-04-16) | 64.9 | #183 |
| 5 | Gemini 1.5 Pro (002) | 58.28 | #368 |
| 6 | GPT-4.1 (2025-04-14) | 58.28 | #257 |
| 7 | GPT-4.5 | 53.64 | #278 |
| 8 | Claude 3.5 Sonnet (20241022) | 51.66 | #302 |
| 9 | Gemini 2.0 Flash | 51.66 | #351 |
| 10 | Claude 3.5 Sonnet (20240620) | 50.99 | #308 |
| 11 | GPT-4o (2024-08-06) | 50.33 | #343 |
| 12 | GPT-4o (2024-11-20) | 50.33 | #392 |
| 13 | GPT-4o (2024-05-13) | 49.67 | #337 |
| 14 | Gemini 2.0 Pro (Preview 02-05) | 49.67 | #318 |
| 15 | Gemini 1.5 Flash (002) | 49.01 | #587 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-vhelm-math-vista-elementary-school-free-form · How It Works · Data refreshed daily, snapshot 2026-11-05.