OpenVLM AI2D - Water, Carbon, and Nitrogen Cycle — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 284 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3-78B | 77.3 |
| 2 | InternVL3-38B | 77.3 |
| 3 | GPT-5 (2025-08-07) | 77.3 |
| 4 | Gemini 2.5 Pro | 75 |
| 5 | InternVL3-14B | 75 |
| 6 | InternVL2.5-78B | 75 |
| 7 | GPT-4.1 (2025-04-14) | 75 |
| 8 | InternVL3-8B | 72.7 |
| 9 | Qwen 2 VL 72B | 72.7 |
| 10 | Claude 3.5 Sonnet (20241022) | 72.7 |
| 11 | Ovis2-8B | 72.7 |
| 12 | Claude 3.5 Sonnet | 72.7 |
| 13 | GPT-4.5 | 72.7 |
| 14 | GPT-4o ChatGPT | 70.5 |
| 15 | InternVL2-8B | 68.2 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-ai2d-water-carbon-and-nitrogen-cycle · How the rankings work · Data refreshed daily, snapshot 2026-07-22.