GeoMMBench - Applications: leaderboard
The remote sensing application 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 remote sensing application 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 | GPT-4o | 72 | #751 |
| 2 | Qwen 3 VL 8B | 71.2 | #917 |
| 3 | GPT-4V | 69.5 | #1470 |
| 4 | Qwen 3 VL 30B | 67.8 | #561 |
| 5 | Gemini 2.0 Flash | 64.4 | #723 |
| 6 | Qwen 2.5 VL 7B | 64.4 | #1862 |
| 7 | O1 | 64.4 | #403 |
| 8 | InternVL2.5-8B | 64.4 | #2370 |
| 9 | Qwen 3 VL 4B | 64.4 | #1151 |
| 10 | Gemini 1.5 Pro | 62.7 | #1063 |
| 11 | Qwen 2 VL 7B | 61.9 | #2275 |
| 12 | GPT-5 | 59.3 | #178 |
| 13 | Phi-4 Multimodal Instruct | 58.5 | #4154 |
| 14 | Qwen 2.5 VL 3B | 57.6 | #2639 |
| 15 | NVILA-15B | 57.6 | #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-applications · How It Works · Data refreshed daily, snapshot 2026-11-21.