# Self-Compression of Chain-of-Thought via Multi-Agent Reinforcement Learning

arXiv: https://arxiv.org/abs/2601.21919 · HTML: https://mlsanity.com/paper/2601.21919/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 38 of its 45 references that are in the map.
- First posted on arXiv 29 Jan 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 30 of 38)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society](https://mlsanity.com/paper/2303.17760/) (2023)
- [s1: Simple test-time scaling](https://mlsanity.com/paper/2501.19393/) (2025)
- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [The Surprising Effectiveness of PPO in Cooperative, Multi-Agent Games](https://mlsanity.com/paper/2103.01955/) (2021)
- [Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs](https://mlsanity.com/paper/2412.21187/) (2024)
- [L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning](https://mlsanity.com/paper/2503.04697/) (2025)
- [Chain of Draft: Thinking Faster by Writing Less](https://mlsanity.com/paper/2502.18600/) (2025)
- [O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning](https://mlsanity.com/paper/2501.12570/) (2025)
- [Token-Budget-Aware LLM Reasoning](https://mlsanity.com/paper/2412.18547/) (2024)
- [Training Language Models to Reason Efficiently](https://mlsanity.com/paper/2502.04463/) (2025)
- [ThinkPrune: Pruning Long Chain-of-Thought of LLMs via Reinforcement Learning](https://mlsanity.com/paper/2504.01296/) (2025)
- [Reasoning Models Can Be Effective Without Thinking](https://mlsanity.com/paper/2504.09858/) (2025)
- [The Danger of Overthinking: Examining the Reasoning-Action Dilemma in Agentic Tasks](https://mlsanity.com/paper/2502.08235/) (2025)
- [AdaptThink: Reasoning Models Can Learn When to Think](https://mlsanity.com/paper/2505.13417/) (2025)
- [Group Robust Preference Optimization in Reward-free RLHF](https://mlsanity.com/paper/2405.20304/) (2024)
- [S-GRPO: Early Exit via Reinforcement Learning in Reasoning Models](https://mlsanity.com/paper/2505.07686/) (2025)
- [GRPO-LEAD: A Difficulty-Aware Reinforcement Learning Approach for Concise Mathematical Reasoning in Language Models](https://mlsanity.com/paper/2504.09696/) (2025)
- [Think Only When You Need with Large Hybrid-Reasoning Models](https://mlsanity.com/paper/2505.14631/) (2025)
- [Learning When to Think: Shaping Adaptive Reasoning in R1-Style Models via Multi-Stage RL](https://mlsanity.com/paper/2505.10832/) (2025)
- [Answer Convergence as a Signal for Early Stopping in Reasoning](https://mlsanity.com/paper/2506.02536/) (2025)
- [Optimizing Length Compression in Large Reasoning Models](https://mlsanity.com/paper/2506.14755/) (2025)
- [Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Models](https://mlsanity.com/paper/2508.02120/) (2025)
- [ConCISE: Confidence-guided Compression in Step-by-step Efficient Reasoning](https://mlsanity.com/paper/2505.04881/) (2025)

## Built on by (most cited first, 13 of 13)

- [Auditing Multi-Agent LLM Reasoning Trees Outperforms Majority Vote and LLM-as-Judge](https://mlsanity.com/paper/2602.09341/) (2026)
- [POET: Power-Oriented Evolutionary Tuning for LLM-Based RTL PPA Optimization](https://mlsanity.com/paper/2603.19333/) (2026)
- [Adaptive Collaboration with Humans: Metacognitive Policy Optimization for Multi-Agent LLMs with Continual Learning](https://mlsanity.com/paper/2603.07972/) (2026)
- [RaMem: Contextual Reinstatement for Long-term Agentic Memory](https://mlsanity.com/paper/2606.22844/) (2026)
- [De-conflating Preference and Qualification: Constrained Dual-Perspective Reasoning for Job Recommendation with Large Language Models](https://mlsanity.com/paper/2602.03097/) (2026)
- [ExpThink: Experience-Guided Reinforcement Learning for Adaptive Chain-of-Thought Compression](https://mlsanity.com/paper/2605.07501/) (2026)
- [SynCo: Data Synthesis Co-Training for Self-Evolving LLMs via Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2610.11345/) (2026)
- [VLM4Rec: Multimodal Semantic Representation for Recommendation with Large Vision-Language Models](https://mlsanity.com/paper/2603.12625/) (2026)
- [TimeMM: Time-as-Operator Spectral Filtering for Dynamic Multimodal Recommendation](https://mlsanity.com/paper/2604.26247/) (2026)
- [GTA: Graph Theory Agent and Benchmark for Algorithmic Graph Reasoning with LLMs](https://mlsanity.com/paper/2609.12265/) (2026)
- [UnifiedPlayers: Enhance Tool-Integrated Reasoning in Agentic Reinforcement Learning](https://mlsanity.com/paper/2609.20089/) (2026)
- [Controlled Decoding Attacks on Black-Box LLMs](https://mlsanity.com/paper/2609.36956/) (2026)
- [Beyond Final Accuracy: Auditing Communication in LLM Multi-Agent Systems](https://mlsanity.com/paper/2610.01042/) (2026)

## Abstract (arXiv)

The inference overhead induced by redundant reasoning undermines the interactive experience and severely bottlenecks the deployment of Large Reasoning Models. Existing reinforcement learning (RL)-based solutions tackle this problem by coupling a length penalty with outcome-based rewards. This simplistic reward weighting struggles to reconcile brevity with accuracy, as enforcing brevity may compromise critical reasoning logic. In this work, we address this limitation by proposing a multi-agent RL framework that selectively penalizes redundant chunks, while preserving essential reasoning logic. Our framework, Self-Compression via MARL (SCMA), instantiates redundancy detection and evaluation through two specialized agents: a Segmentation Agent for decomposing the reasoning process into logical chunks, and a Scoring Agent for quantifying the significance of each chunk. The Segmentation and Scoring agents collaboratively define an importance-weighted length penalty during training, incentivizing a Reasoning Agent to prioritize essential logic without introducing inference overhead during deployment. Empirical evaluations across model scales demonstrate that SCMA reduces response length by 11.1% to 39.0% while boosting accuracy by 4.33% to 10.02%. Furthermore, ablation studies and qualitative analysis validate that the synergistic optimization within the MARL framework fosters emergent behaviors, yielding more powerful LRMs compared to vanilla RL paradigms.