CPMobius: Iterative Coach-Player Reasoning for Data-Free Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Ran, Liu, Zeyuan, Chen, Yinghao, He, Bingxiang, Yuan, Jiarui, Fu, Zixuan, Chen, Weize, Hu, Jinyi, Qian, Chen, Liu, Zhiyuan, Sun, Maosong
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913160891465728
author Li, Ran
Liu, Zeyuan
Chen, Yinghao
He, Bingxiang
Yuan, Jiarui
Fu, Zixuan
Chen, Weize
Hu, Jinyi
Qian, Chen
Liu, Zhiyuan
Sun, Maosong
author_facet Li, Ran
Liu, Zeyuan
Chen, Yinghao
He, Bingxiang
Yuan, Jiarui
Fu, Zixuan
Chen, Weize
Hu, Jinyi
Qian, Chen
Liu, Zhiyuan
Sun, Maosong
contents Large Language Models (LLMs) have demonstrated strong potential in complex reasoning, yet their progress remains fundamentally constrained by reliance on massive high-quality human-curated tasks and labels, either through supervised fine-tuning (SFT) or reinforcement learning (RL) on reasoning-specific data. This dependence renders supervision-heavy training paradigms increasingly unsustainable, with signs of diminishing scalability already evident in practice. To overcome this limitation, we introduce CPMöbius (CPMobius), a collaborative Coach-Player paradigm for data-free reinforcement learning of reasoning models. Unlike traditional adversarial self-play, CPMöbius, inspired by real world human sports collaboration and multi-agent collaboration, treats the Coach and Player as independent but cooperative roles. The Coach proposes instructions targeted at the Player's capability and receives rewards based on changes in the Player's performance, while the Player is rewarded for solving the increasingly instructive tasks generated by the Coach. This cooperative optimization loop is designed to directly enhance the Player's mathematical reasoning ability. Remarkably, CPMöbius achieves substantial improvement without relying on any external training data, outperforming existing unsupervised approaches. For example, on Qwen2.5-Math-7B-Instruct, our method improves accuracy by an overall average of +4.9 and an out-of-distribution average of +5.4, exceeding RENT by +1.5 on overall accuracy and R-zero by +4.2 on OOD accuracy. Our codebase has been released at https://github.com/thunlp/CPMobius.
format Preprint
id arxiv_https___arxiv_org_abs_2602_02979
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CPMobius: Iterative Coach-Player Reasoning for Data-Free Reinforcement Learning
Li, Ran
Liu, Zeyuan
Chen, Yinghao
He, Bingxiang
Yuan, Jiarui
Fu, Zixuan
Chen, Weize
Hu, Jinyi
Qian, Chen
Liu, Zhiyuan
Sun, Maosong
Computation and Language
Machine Learning
Large Language Models (LLMs) have demonstrated strong potential in complex reasoning, yet their progress remains fundamentally constrained by reliance on massive high-quality human-curated tasks and labels, either through supervised fine-tuning (SFT) or reinforcement learning (RL) on reasoning-specific data. This dependence renders supervision-heavy training paradigms increasingly unsustainable, with signs of diminishing scalability already evident in practice. To overcome this limitation, we introduce CPMöbius (CPMobius), a collaborative Coach-Player paradigm for data-free reinforcement learning of reasoning models. Unlike traditional adversarial self-play, CPMöbius, inspired by real world human sports collaboration and multi-agent collaboration, treats the Coach and Player as independent but cooperative roles. The Coach proposes instructions targeted at the Player's capability and receives rewards based on changes in the Player's performance, while the Player is rewarded for solving the increasingly instructive tasks generated by the Coach. This cooperative optimization loop is designed to directly enhance the Player's mathematical reasoning ability. Remarkably, CPMöbius achieves substantial improvement without relying on any external training data, outperforming existing unsupervised approaches. For example, on Qwen2.5-Math-7B-Instruct, our method improves accuracy by an overall average of +4.9 and an out-of-distribution average of +5.4, exceeding RENT by +1.5 on overall accuracy and R-zero by +4.2 on OOD accuracy. Our codebase has been released at https://github.com/thunlp/CPMobius.
title CPMobius: Iterative Coach-Player Reasoning for Data-Free Reinforcement Learning
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2602.02979