VMMU - Chemistry: leaderboard
Metric: Accuracy (%) on the VMMU Chemistry questions (Vietnamese page-image multiple choice); higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro | 91.72 |
| 2 | GPT-5 (Medium) | 85.1 |
| 3 | O3 (Medium) | 82.78 |
| 4 | Claude Opus 4.6 (Thinking) | 79.47 |
| 5 | Gemini 2.5 Flash | 78.15 |
| 6 | Claude Sonnet 4 | 53.31 |
| 7 | Qwen 2.5 VL 72B Instruct | 48.34 |
| 8 | Llama 4 Scout | 45.03 |
| 9 | Qwen 2.5 VL 32B Instruct | 44.7 |
| 10 | GPT-4.1 | 44.37 |
| 11 | Gemma 3 27B | 38.74 |
| 12 | Mistral Medium 3 | 37.75 |
| 13 | Llama 4 Maverick | 32.78 |
| 14 | Mistral Small 3.2 | 28.48 |
| 15 | aya-vision-8B | 27.81 |
Interactive version: theaggregate.ai/benchmark?slug=vmmu-chemistry · How It Works · Data refreshed daily, snapshot 2026-09-29.