MMBU - Grounded Classification (Boxes): leaderboard
Metric: Closed-ended classification of a region marked by a bounding box (from detection annotations): micro-averaged F1 x100 (0-100) over MMBU's biomedical images from 35 submodalities (95% bootstrap intervals not published); zero-shot, official checkpoints; exact match after normalization. Source: arxiv.org. Saturation forecast: Around July 2028. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3.5-8B | 43.9 |
| 2 | Qwen 3 VL 32B Instruct | 43.3 |
| 3 | GPT-4.1 Mini | 43 |
| 4 | GPT-5.4 Mini | 41 |
| 5 | Qwen 2.5 VL 32B Instruct | 40.1 |
| 6 | Qwen 3 VL 8B Instruct | 39.1 |
| 7 | MedGemma-4B-IT | 37.1 |
| 8 | Lingshu-32B | 36.8 |
| 9 | Qwen 3 VL 4B Instruct | 36.3 |
| 10 | Lingshu-7B | 36 |
| 11 | Qwen 2.5 VL 7B Instruct | 32 |
| 12 | Gemma 3 4B (IT) | 28.3 |
Interactive version: theaggregate.ai/benchmark?slug=mmbu-grounded-classification-boxes · How It Works · Data refreshed daily, snapshot 2026-09-29.