OpenVLM AI2D - Eclipses: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 284 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 100 |
| 2 | InternVL3-78B | 100 |
| 3 | Qwen 2 VL 72B | 100 |
| 4 | InternVL3-38B | 96.8 |
| 5 | InternVL2.5-78B | 96.8 |
| 6 | GPT-5 | 93.5 |
| 7 | InternVL3-8B | 93.5 |
| 8 | Aquila-VL-2B | 93.5 |
| 9 | Gemini 1.5 Flash (002) | 93.5 |
| 10 | Grok 2 (1212) | 93.5 |
| 11 | GPT-4.1 (2025-04-14) | 93.5 |
| 12 | GPT-4o ChatGPT | 93.5 |
| 13 | InternVL2-8B | 90.3 |
| 14 | Qwen 2 VL 7B | 90.3 |
| 15 | Claude 3.5 Sonnet (20241022) | 90.3 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-ai2d-eclipses · How It Works · Data refreshed daily, snapshot 2026-09-05.