OpenVLM MMT-Bench - Raven Progressive Matrices — leaderboard
Metric: Score (%). Source: huggingface.co. 207 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2 VL 72B | 65 |
| 2 | Claude 3.5 Sonnet | 35 |
| 3 | Gemma 3 12B | 25 |
| 4 | GPT-4.1 Mini | 20 |
| 5 | InternVL3-78B | 20 |
| 6 | InternVL2-8B | 20 |
| 7 | Gemini 1.5 Pro | 20 |
| 8 | GPT-4.1 (2025-04-14) | 20 |
| 9 | InternVL3-8B | 15 |
| 10 | Qwen 2 VL 7B | 15 |
| 11 | InternVL3-38B | 15 |
| 12 | InternVL3-14B | 15 |
| 13 | Aquila-VL-2B | 15 |
| 14 | MiniCPM-V-2.6 | 15 |
| 15 | GPT-4.1 Nano | 15 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-raven-progressive-matrices · How the rankings work · Data refreshed daily, snapshot 2026-07-22.