OpenVLM MME - Count: leaderboard
OpenCompass OpenVLM MME subcategory score for counting objects in visual inputs.
Metric: Score. Source: huggingface.co. Saturation forecast: Estimated already saturated. 235 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 (2025-04-14) | 173.3 | #250 |
| 2 | InternVL3-78B | 170 | #352 |
| 3 | Qwen 2 VL 72B | 170 | #460 |
| 4 | InternVL2.5-78B | 170 | #431 |
| 5 | GPT-4.1 Mini | 165 | #355 |
| 6 | Gemini 2.0 Flash | 165 | #342 |
| 7 | InternVL3-38B | 165 | #405 |
| 8 | InternVL2-8B | 165 | #817 |
| 9 | Qwen 2.5 VL 7B | 165 | #643 |
| 10 | Grok 2 (1212) | 165 | #496 |
| 11 | Moonshot v1 8K | 163.3 | #617 |
| 12 | InternVL3-8B | 160 | #620 |
| 13 | MiniCPM-V-2.6 | 160 | #836 |
| 14 | Qwen 2 VL 7B | 160 | #750 |
| 15 | InternVL3-14B | 160 | #505 |
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-count · How It Works · Data refreshed daily, snapshot 2026-10-20.