OpenVLM AI2D - Water, Carbon, and Nitrogen Cycle: leaderboard

OpenCompass OpenVLM AI2D subcategory score for diagram questions about water, carbon, and nitrogen cycles.

Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 284 models tracked.

Top models

#ModelScoreOverall rank
1GPT-577.3#98
2InternVL3-38B77.3#405
3InternVL3-78B77.3#352
4Gemini 2.5 Pro75#152
5GPT-4.1 (2025-04-14)75#250
6InternVL3-14B75#505
7InternVL2.5-78B75#431
8Claude 3.5 Sonnet (20241022)72.7#291
9InternVL3-8B72.7#620
10Qwen 2 VL 72B72.7#460
11Claude 3.5 Sonnet72.7#350
12GPT-4.572.7#274
13GPT-4o ChatGPT70.5
14GPT-5 Nano68.2#425
15InternVL2-8B68.2#817

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-water-carbon-and-nitrogen-cycle · How It Works · Data refreshed daily, snapshot 2026-10-20.