OpenVLM AI2D - Faults and Earthquakes: leaderboard
OpenCompass OpenVLM AI2D subcategory score for diagram questions about faults and earthquakes.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around January 2027. 284 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL2.5-78B | 92.9 | #431 |
| 2 | Gemini 2.5 Pro | 89.3 | #152 |
| 3 | Claude 3.7 Sonnet | 85.7 | #249 |
| 4 | Qwen 2 VL 72B | 85.7 | #460 |
| 5 | InternVL3-38B | 82.1 | #405 |
| 6 | InternVL3-78B | 82.1 | #352 |
| 7 | GPT-5 | 78.6 | #98 |
| 8 | InternVL3-8B | 78.6 | #620 |
| 9 | Gemini 1.5 Flash (002) | 78.6 | #583 |
| 10 | InternVL3-14B | 78.6 | #505 |
| 11 | GPT-5 Mini (2025-08-07) | 78.6 | #172 |
| 12 | GPT-4.5 | 78.6 | #274 |
| 13 | GPT-5 Nano | 71.4 | #425 |
| 14 | InternVL2-8B | 71.4 | #817 |
| 15 | MiniCPM-V-2.6 | 71.4 | #836 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openvlm-ai2d-faults-and-earthquakes · How It Works · Data refreshed daily, snapshot 2026-10-20.