OpenVLM MME - Cognition: leaderboard
OpenCompass OpenVLM MME subcategory score for visual cognition and reasoning questions.
Metric: Score. Source: huggingface.co. Saturation forecast: Estimated already saturated. 235 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-78B | 753.2 | #352 |
| 2 | Qwen 2 VL 72B | 733.6 | #460 |
| 3 | InternVL3-38B | 727.1 | #405 |
| 4 | InternVL3-14B | 720.7 | #505 |
| 5 | InternVL2.5-78B | 719.3 | #431 |
| 6 | Gemini 2.0 Flash | 707.5 | #342 |
| 7 | GPT-4.1 (2025-04-14) | 673.9 | #250 |
| 8 | InternVL3-8B | 673.6 | #620 |
| 9 | Claude 3.7 Sonnet | 649.6 | #249 |
| 10 | Qwen 2 VL 7B | 637.1 | #750 |
| 11 | Moonshot v1 8K | 625.7 | #617 |
| 12 | GPT-4.1 Mini | 615.7 | #355 |
| 13 | Qwen 2.5 VL 7B | 613.9 | #643 |
| 14 | Claude 3.5 Sonnet (20241022) | 599.6 | #291 |
| 15 | MiniCPM-V-2.6 | 597.9 | #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-mme-cognition · How It Works · Data refreshed daily, snapshot 2026-10-20.