OpenVLM MMT-Bench - Whoops: leaderboard
The "Whoops" 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. 207 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-78B | 95 | #352 |
| 2 | InternVL3-14B | 95 | #505 |
| 3 | Qwen 2 VL 2B | 95 | #1220 |
| 4 | Qwen 2 VL 72B | 90 | #460 |
| 5 | Gemini 1.5 Flash | 90 | #618 |
| 6 | GPT-4.1 Mini | 85 | #355 |
| 7 | InternVL3-38B | 85 | #405 |
| 8 | MiniCPM-V-2.6 | 85 | #836 |
| 9 | Claude 3.5 Sonnet | 85 | #350 |
| 10 | Qwen 2.5 VL 7B | 85 | #643 |
| 11 | Claude 3 Haiku | 85 | #803 |
| 12 | Claude 3 Sonnet | 85 | #775 |
| 13 | InternVL2-8B | 80 | #817 |
| 14 | Qwen 2 VL 7B | 80 | #750 |
| 15 | Aquila-VL-2B | 80 | #1238 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-whoops · How It Works · Data refreshed daily, snapshot 2026-10-20.