Saved in:
Bibliographic Details
Main Authors: M3 Team, Dou, Chengfeng, Yang, Fan, Li, Fei, Jia, Jiyuan, Ju, Qiang, Wang, Shuai, Li, Tianpeng, Zeng, Xiangrong, Zhou, Yijie, Zhang, Hongda, Tai, Jinyang, Sun, Linzhuang, Guo, Peidong, Mo, Yichuan, Wang, Xiaochuan, Cui, Hengfu, Zhang, Zhishou
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.06570
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910014107549696
author M3 Team
Dou, Chengfeng
Yang, Fan
Li, Fei
Jia, Jiyuan
Ju, Qiang
Wang, Shuai
Li, Tianpeng
Zeng, Xiangrong
Zhou, Yijie
Zhang, Hongda
Tai, Jinyang
Sun, Linzhuang
Guo, Peidong
Mo, Yichuan
Wang, Xiaochuan
Cui, Hengfu
Zhang, Zhishou
author_facet M3 Team
Dou, Chengfeng
Yang, Fan
Li, Fei
Jia, Jiyuan
Ju, Qiang
Wang, Shuai
Li, Tianpeng
Zeng, Xiangrong
Zhou, Yijie
Zhang, Hongda
Tai, Jinyang
Sun, Linzhuang
Guo, Peidong
Mo, Yichuan
Wang, Xiaochuan
Cui, Hengfu
Zhang, Zhishou
contents We introduce Baichuan-M3, a medical-enhanced large language model engineered to shift the paradigm from passive question-answering to active, clinical-grade decision support. Addressing the limitations of existing systems in open-ended consultations, Baichuan-M3 utilizes a specialized training pipeline to model the systematic workflow of a physician. Key capabilities include: (i) proactive information acquisition to resolve ambiguity; (ii) long-horizon reasoning that unifies scattered evidence into coherent diagnoses; and (iii) adaptive hallucination suppression to ensure factual reliability. Empirical evaluations demonstrate that Baichuan-M3 achieves state-of-the-art results on HealthBench, the newly introduced HealthBench-Hallu and ScanBench, significantly outperforming GPT-5.2 in clinical inquiry, advisory and safety. The models are publicly available at https://huggingface.co/collections/baichuan-inc/baichuan-m3.
format Preprint
id arxiv_https___arxiv_org_abs_2602_06570
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Baichuan-M3: Modeling Clinical Inquiry for Reliable Medical Decision-Making
M3 Team
Dou, Chengfeng
Yang, Fan
Li, Fei
Jia, Jiyuan
Ju, Qiang
Wang, Shuai
Li, Tianpeng
Zeng, Xiangrong
Zhou, Yijie
Zhang, Hongda
Tai, Jinyang
Sun, Linzhuang
Guo, Peidong
Mo, Yichuan
Wang, Xiaochuan
Cui, Hengfu
Zhang, Zhishou
Computation and Language
We introduce Baichuan-M3, a medical-enhanced large language model engineered to shift the paradigm from passive question-answering to active, clinical-grade decision support. Addressing the limitations of existing systems in open-ended consultations, Baichuan-M3 utilizes a specialized training pipeline to model the systematic workflow of a physician. Key capabilities include: (i) proactive information acquisition to resolve ambiguity; (ii) long-horizon reasoning that unifies scattered evidence into coherent diagnoses; and (iii) adaptive hallucination suppression to ensure factual reliability. Empirical evaluations demonstrate that Baichuan-M3 achieves state-of-the-art results on HealthBench, the newly introduced HealthBench-Hallu and ScanBench, significantly outperforming GPT-5.2 in clinical inquiry, advisory and safety. The models are publicly available at https://huggingface.co/collections/baichuan-inc/baichuan-m3.
title Baichuan-M3: Modeling Clinical Inquiry for Reliable Medical Decision-Making
topic Computation and Language
url https://arxiv.org/abs/2602.06570