• Model confrontation and collab...

*Result*: Model confrontation and collaboration: A debate intelligence framework for enhancing medical reasoning in large language models.

Title:
Model confrontation and collaboration: A debate intelligence framework for enhancing medical reasoning in large language models.
Authors:
Sun X; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Hong Q; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Zhang M; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Li Y; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Chen T; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Huang Z; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Liang G; School of Basic Medical Sciences, Peking University, Beijing, China., Tang W; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Xu S; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Ni X; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Pang J; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China., Wan P; School of Basic Medical Sciences, Peking University, Beijing, China. Electronic address: peixing@bjmu.edu.cn., Long E; State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China. Electronic address: erping.long@ibms.pumc.edu.cn.
Source:
Cell reports. Medicine [Cell Rep Med] 2026 Jan 20; Vol. 7 (1), pp. 102547. Date of Electronic Publication: 2026 Jan 05.
Publication Type:
Journal Article
Language:
English
Journal Info:
Publisher: Cell Press Country of Publication: United States NLM ID: 101766894 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 2666-3791 (Electronic) Linking ISSN: 26663791 NLM ISO Abbreviation: Cell Rep Med Subsets: MEDLINE
Imprint Name(s):
Original Publication: [Cambridge, MA] : Cell Press, [2020]-
MeSH Terms:
Clinical Decision-Making*/methods , Cooperative Behavior* , Language* , Clinical Reasoning* , Models, Theoretical*, Humans ; Communication ; Large Language Models
References:
Find ACL EMNLP. 2024 Nov;2024:12448-12465. (PMID: 41341631)
Nat Commun. 2025 Jan 14;16(1):642. (PMID: 39809759)
BMC Med Inform Decis Mak. 2020 Nov 30;20(1):310. (PMID: 33256715)
Nat Med. 2024 Oct;30(10):2878-2885. (PMID: 39009780)
Comput Biol Med. 2023 Sep;163:107096. (PMID: 37302375)
NPJ Digit Med. 2024 Jul 23;7(1):190. (PMID: 39043988)
Nat Med. 2025 Mar;31(3):943-950. (PMID: 39779926)
Nature. 2023 Aug;620(7972):172-180. (PMID: 37438534)
Nat Biomed Eng. 2025 Apr;9(4):432-438. (PMID: 40169759)
Stud Health Technol Inform. 2019 Aug 21;264:25-29. (PMID: 31437878)
Nat Commun. 2025 Oct 23;16(1):9377. (PMID: 41130954)
Nature. 2025 Jun;642(8067):442-450. (PMID: 40205050)
Nature. 2025 Sep;645(8081):633-638. (PMID: 40962978)
Nat Med. 2025 Feb;31(2):599-608. (PMID: 39511432)
NPJ Digit Med. 2025 Mar 13;8(1):159. (PMID: 40082662)
Nat Med. 2024 Dec;30(12):3590-3600. (PMID: 39313595)
Contributed Indexing:
Keywords: large language models; medical AI benchmarks; medical reasoning
Entry Date(s):
Date Created: 20260106 Date Completed: 20260121 Latest Revision: 20260206
Update Code:
20260206
PubMed Central ID:
PMC12866169
DOI:
10.1016/j.xcrm.2025.102547
PMID:
41494532
Database:
MEDLINE

*Further Information*

*Medical reasoning is fundamental to clinical decision-making, underpinning tasks such as patient communication, diagnosis, and treatment planning. Inspired by psychological findings that peer interaction promotes self-correction, we introduce model confrontation and collaboration (MCC), a debate intelligence framework that transcends static ensemble methods by integrating critique and self-reflection to iteratively refine reasoning through structured, multi-round confrontation and collaboration among diverse large language models (LLMs). In multiple-choice benchmarks, MCC achieved mean accuracy on MedQA (92.6%) and PubMedQA (84.8%) and demonstrated strong performance on medical subsets of MMLU. In long-form medical question answering, MCC outperformed all individual LLMs and the domain-specific LLM Med-PaLM 2 in both physician and layperson evaluations. In diagnostic dialog tasks, MCC further excelled in both history-taking and diagnostic accuracy, reaching a top-1 diagnosis rate of 80%. These results position MCC as a scalable, model-agnostic framework that advances medical reasoning through collaborative deliberation.
(Copyright © 2025 The Author(s). Published by Elsevier Inc. All rights reserved.)*

*Declaration of interests The authors declare no competing interests.*