GeoMMBench - DEM: leaderboard
The digital elevation model questions of GeoMMBench (2026), an expert-written multimodal geoscience and remote sensing question benchmark answered zero-shot (test set). Micro-averaged accuracy in percent; higher is better.
Metric: Micro-averaged accuracy (%) on the digital elevation model questions of the GeoMMBench test set (expert-written multimodal geoscience and remote sensing multiple-choice and numeric questions, zero-shot, LMMs-Eval, rule-based answer extraction, invalid outputs counted wrong); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 35 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.0 Flash | 67.1 | #723 |
| 2 | Gemini 1.5 Pro | 64.7 | #1063 |
| 3 | O1 | 63.5 | #403 |
| 4 | GPT-4o | 62.3 | #751 |
| 5 | GPT-4V | 61.7 | #1470 |
| 6 | Qwen 3 VL 30B | 59.9 | #561 |
| 7 | Qwen 3 VL 8B | 59.3 | #917 |
| 8 | Qwen 2.5 VL 7B | 52.1 | #1862 |
| 9 | Qwen 2 VL 7B | 51.5 | #2275 |
| 10 | Phi-4 Multimodal Instruct | 50.3 | #4154 |
| 11 | LLaVA-OneVision-7B | 50.3 | #2981 |
| 12 | Qwen 3 VL 4B | 49.7 | #1151 |
| 13 | Qwen 2.5 VL 3B | 48.5 | #2639 |
| 14 | CogVLM2 | 48.5 | |
| 15 | NVILA-15B | 47.3 | #1929 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=geommbench-dem · How It Works · Data refreshed daily, snapshot 2026-11-21.