MMGist - STEM Reasoning: leaderboard
Metric: Accuracy (%) on the STEM Reasoning dimension (items kept from MathVista, MathVision, DynaMath and MME-SCI); MMGist: 7,262 curated items from 18 vision-language benchmarks after removing items answerable without the image, items nearly every model solves and items with faulty labels; eight samples per item at temperature 1.0, step-by-step answer in a boxed block, 16,384 output tokens, medium effort where configurable; higher is better. Source: arxiv.org. Saturation forecast: Around October 2027. 27 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.6 Plus | 62.6 |
| 2 | Gemini 3.1 Pro (Preview) (Medium) | 61.2 |
| 3 | Seed 2.0 Pro | 57.6 |
| 4 | Qwen 3.6 35B A3B | 54.3 |
| 5 | Seed 2.0 Lite | 53.1 |
| 6 | Seed 2.0 Mini | 48.3 |
| 7 | Qwen 3.5 27B | 46.6 |
| 8 | GPT-5 (Medium) | 46.1 |
| 9 | Claude Sonnet 4.6 (Medium) | 44.6 |
| 10 | Gemini 3.1 Flash Lite (Medium) | 41.5 |
| 11 | Gemma 4 31B | 40.4 |
| 12 | Qwen 3.5 9B | 37.1 |
| 13 | Gemma 4 26B A4B | 36.8 |
| 14 | GPT-5 Mini (Medium) | 36.8 |
| 15 | Step3 VL 10B | 36.3 |
Interactive version: theaggregate.ai/benchmark?slug=mmgist-stem-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-29.