Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.00842 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910036728479744 |
|---|---|
| author | Zhang, Kai Yuan, Zhengqing Peng, Cheng Zhao, Songlin Lyu, Mengxian Chen, Ziyi Ye, Yanfang Liu, Wei Zhang, Ying Smith, Kaleb E He, Lifang Sun, Lichao Wu, Yonghui |
| author_facet | Zhang, Kai Yuan, Zhengqing Peng, Cheng Zhao, Songlin Lyu, Mengxian Chen, Ziyi Ye, Yanfang Liu, Wei Zhang, Ying Smith, Kaleb E He, Lifang Sun, Lichao Wu, Yonghui |
| contents | Biomedical multimodal assistants have the potential to unify radiology, pathology, and clinical-text reasoning, yet a critical deployment gap remains: top-performing systems are either closed-source or computationally prohibitive, precluding the on-premises deployment required for patient privacy and PHI compliance. We introduce MEDGPT-OSS, an open-weight, 20B-parameter generalist vision-language model designed to facilitate open research in clinical AI. Rather than relying on architectural complexity, MEDGPT-OSS pairs the GPT-oss language backbone with a visual front-end via a optimized, three-stage training curriculum. By progressively domain-adapting these modules through rigorous data curation and long-context multimodal alignment, we demonstrate that a 20B model can bridge the capacity gap. It successfully outperforms larger open medical models on out-of-distribution (OOD) multimodal reasoning and complex text-only clinical tasks. By unifying diverse modalities under a single instruction-following interface, MEDGPT-OSS maintains a parameter-efficient footprint fully compatible with commodity GPUs. We release the complete training recipe, open-weight checkpoints, and a rigorous evaluation harness to serve as a verifiable foundation for privacy-preserving, institution-specific clinical AI research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_00842 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MedGPT-oss: Training a General-Purpose Vision-Language Model for Biomedicine Zhang, Kai Yuan, Zhengqing Peng, Cheng Zhao, Songlin Lyu, Mengxian Chen, Ziyi Ye, Yanfang Liu, Wei Zhang, Ying Smith, Kaleb E He, Lifang Sun, Lichao Wu, Yonghui Computation and Language Biomedical multimodal assistants have the potential to unify radiology, pathology, and clinical-text reasoning, yet a critical deployment gap remains: top-performing systems are either closed-source or computationally prohibitive, precluding the on-premises deployment required for patient privacy and PHI compliance. We introduce MEDGPT-OSS, an open-weight, 20B-parameter generalist vision-language model designed to facilitate open research in clinical AI. Rather than relying on architectural complexity, MEDGPT-OSS pairs the GPT-oss language backbone with a visual front-end via a optimized, three-stage training curriculum. By progressively domain-adapting these modules through rigorous data curation and long-context multimodal alignment, we demonstrate that a 20B model can bridge the capacity gap. It successfully outperforms larger open medical models on out-of-distribution (OOD) multimodal reasoning and complex text-only clinical tasks. By unifying diverse modalities under a single instruction-following interface, MEDGPT-OSS maintains a parameter-efficient footprint fully compatible with commodity GPUs. We release the complete training recipe, open-weight checkpoints, and a rigorous evaluation harness to serve as a verifiable foundation for privacy-preserving, institution-specific clinical AI research. |
| title | MedGPT-oss: Training a General-Purpose Vision-Language Model for Biomedicine |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2603.00842 |