OpenVLM MMT-Bench - Count: leaderboard
The "Count" subtask of OpenVLM MMT-Bench. OpenCompass OpenVLM evaluation of MMT-Bench: multimodal tasks covering visual recognition, localization, OCR, counting, relation reasoning, visual math, and hallucination detection.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 207 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2 VL 72B | 59.1 | #460 |
| 2 | MiniCPM-V-2.6 | 58.1 | #836 |
| 3 | InternVL3-14B | 55.9 | #505 |
| 4 | GPT-4.1 Mini | 55 | #355 |
| 5 | InternVL3-8B | 55 | #620 |
| 6 | Grok 2 (1212) | 55 | #496 |
| 7 | InternVL3-38B | 52.8 | #405 |
| 8 | GPT-4.1 (2025-04-14) | 52.8 | #250 |
| 9 | InternVL3-78B | 51.6 | #352 |
| 10 | Claude 3.5 Sonnet | 50.9 | #350 |
| 11 | Qwen 2.5 VL 7B | 48.4 | #643 |
| 12 | Qwen 2 VL 2B | 47.8 | #1220 |
| 13 | InternVL2-8B | 47.2 | #817 |
| 14 | Pixtral-12B | 46.6 | #808 |
| 15 | Claude 3.7 Sonnet | 46.2 | #249 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-count · How It Works · Data refreshed daily, snapshot 2026-10-20.