Research map / Vision-language and multimodal / Multimodal LLMs
MMBU: A Massive Multi-modal Biomedical Understanding Benchmark to Probe the Perception Capabilities of Vision-Language Models
Alejandro Lozano, Ryan D'Cunha, Daniel Jarquin, Min Sun, Josiah Aklilu, James Burgess, Yuhui Zhang, Paola Robayo, Jin Ye, Ming Hu, Zhongying Deng, Junjun He and 6 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add the image input (vs text-only) (ablation) | OctoMed-7B · MMBU closed VQA | F1: 0.161 -> 0.361 | Improved |
| add the image input (vs text-only) (ablation) | Qwen3-VL-32B · MMBU closed VQA | F1: 0.211 -> 0.405 | Improved |
| add the image input (vs text-only) (ablation) | MedGemma-1.5-4B · MMBU closed VQA | F1: 0.201 -> 0.309 | Improved |
| add the image input (vs text-only) (ablation) | Qwen3-VL-32B · MMBU open VQA | F1: 0.016 -> 0.117 | Improved |
| add the image input (vs text-only) (ablation) | OctoMed-7B · MMBU open VQA | F1: 0.045 -> 0.048 | No clear effect |
| medical adaptation (MedGemma-4B vs Gemma-3-4B) | Gemma-3-4B · MMBU ungrounded classification (closed) | micro-F1: 0.296 -> 0.317 | Improved |
| medical adaptation (LLaVA-Med-7B vs LLaVA-v1.5-7B) | LLaVA-v1.5-7B · MMBU ungrounded classification (closed) | micro-F1: 0.254 -> 0.157 | Hurt |
| medical adaptation (MedGemma-1.5-4B vs Gemma-3-4B) | Gemma-3-4B · MMBU grounded classification from detection (closed) | micro-F1: 0.183 -> 0.204 | Improved |
| medical adaptation (Lingshu-32B vs Qwen2.5-VL-32B) | Qwen2.5-VL-32B · MMBU domain radar area | area: 1.03 -> 0.70 | Hurt |
| medical adaptation (OctoMed-7b vs Qwen2.5-VL-7b) | Qwen2.5-VL-7B · MMBU domain radar area | area: 0.43 -> 0.83 | Improved |
| scale model size 3B to 32B | Qwen2.5-VL · MMBU ungrounded classification (closed) | micro-F1: 0.109 -> 0.320 | Improved |
| scale Lingshu from 7B to 32B | Lingshu · MMBU ungrounded classification (closed) | micro-F1: 0.150 -> 0.333 | Improved |
Closest papers by meaning
- BIOMEDICA: An Open Biomedical Image-Caption Archive, Dataset, and Vision-Language Models Derived from Scientific Literature (CVPR 2025 · similarity 0.90)
- Micro-Bench: A Microscopy Benchmark for Vision-Language Understanding (NeurIPS 2024 · similarity 0.90)
- LLaDA-MedV: Exploring Large Language Diffusion Models for Biomedical Image Understanding (CVPR 2026 · similarity 0.90)
- PathView-Bench: Can Multimodal Large Language Models Achieve Fine-grained Multiscale Understanding of Pathology Images? (NeurIPS 2026 · similarity 0.88)
- BiomedCoOp: Learning to Prompt for Biomedical Vision-Language Models (CVPR 2025 · similarity 0.88)
- How Do Medical MLLMs Fail? A Study on Visual Grounding in Medical Images (ICLR 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
