VMMU - Geography: leaderboard
VMMU's geography questions from Vietnam's national high-school graduation exam, many reading charts, maps and tables, given as rendered page images; accuracy, higher is better.
Metric: Accuracy (%) on the VMMU Geography questions (Vietnamese page-image multiple choice); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 17 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 94.59 | #137 |
| 2 | Claude Opus 4.6 (Thinking) | 90.85 | #102 (Claude Opus 4.6) |
| 3 | GPT-5 (Medium) | 89.4 | #167 (GPT-5) |
| 4 | O3 (Medium) | 88.98 | #228 (O3) |
| 5 | Gemini 2.5 Flash | 85.24 | #453 |
| 6 | Qwen 2.5 VL 32B Instruct | 71.31 | #1053 |
| 7 | GPT-4.1 | 69.85 | #495 |
| 8 | Llama 4 Scout | 69.02 | #1776 |
| 9 | Qwen 2.5 VL 72B Instruct | 64.66 | #846 |
| 10 | Llama 4 Maverick | 53.22 | #1099 |
| 11 | Claude Sonnet 4 | 48.44 | #400 |
| 12 | Gemma 3 27B | 47.4 | #1631 |
| 13 | Mistral Medium 3 | 36.59 | #1073 |
| 14 | Mistral Small 3.2 | 30.98 | #1591 |
| 15 | Gemma 3 4B | 27.65 | #5064 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=vmmu-geography · How It Works · Data refreshed daily, snapshot 2026-11-13.