OpenVLM MME - Perception: leaderboard
OpenCompass OpenVLM MME subcategory score for visual perception tasks.
Metric: Score. Source: huggingface.co. Saturation forecast: Estimated already saturated. 235 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-78B | 1785.3 | #352 |
| 2 | InternVL3-38B | 1773.6 | #405 |
| 3 | Qwen 2 VL 72B | 1770.6 | #460 |
| 4 | InternVL3-8B | 1748.4 | #620 |
| 5 | InternVL2.5-78B | 1748.3 | #431 |
| 6 | InternVL3-14B | 1741.6 | #505 |
| 7 | Qwen 2.5 VL 7B | 1698.1 | #643 |
| 8 | Gemini 2.0 Flash | 1678.6 | #342 |
| 9 | MiniCPM-V-2.6 | 1670.9 | #836 |
| 10 | GPT-4.1 (2025-04-14) | 1663.6 | #250 |
| 11 | Grok 2 (1212) | 1660.8 | #496 |
| 12 | InternVL2-8B | 1639.7 | #817 |
| 13 | Qwen 2 VL 7B | 1639.2 | #750 |
| 14 | Gemini 1.0 Pro | 1608.9 | #933 |
| 15 | Gemini 1.5 Flash | 1589.3 | #618 |
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-perception · How It Works · Data refreshed daily, snapshot 2026-10-20.