← Back to Research
arxiv
LLM February 27, 2026 Impact: 8/10

Towards Efficient Large Language Reasoning Models via Extreme-Ratio Chain-of-Thought Compression

By Yuntian Tang, Bohan Jia, Wenxuan Huang, Lianyue Zhang, Jiao Xie, Wenxi Li, Wei Li, Jie Hu, Xinghao Chen, Rongrong Ji, Shaohui Lin
Chain-of-Thought (CoT) reasoning successfully enhances the reasoning capabilities of Large Language Models (LLMs), yet it incurs substantial computational overhead for inference. Existing CoT compression methods often suffer from a critical loss of logical fidelity at high compression ratios, resulting in significant performance degradation. To achieve high-fidelity, fast reasoning, we propose a novel EXTreme-RAtio Chain-of-Thought Compression framework, termed Extra-CoT, which aggressively reduces the token budget while preserving answer accuracy. To generate reliable, high-fidelity supervision, we first train a dedicated semantically-preserved compressor on mathematical CoT data with fine-grained annotations. An LLM is then fine-tuned on these compressed pairs via a mixed-ratio supervised fine-tuning (SFT), teaching it to follow a spectrum of compression budgets and providing a stable initialization for reinforcement learning (RL). We further propose Constrained and Hierarchical Ratio Policy Optimization (CHRPO) to explicitly incentivize question-solving ability under lower budgets by a hierarchical reward.
chain-of-thoughtreasoninginference efficiencytoken compressionreinforcement learning
Analysis

Summary

Extra-CoT tackles one of the most pressing practical problems in deploying large reasoning models: Chain-of-Thought traces are computationally expensive at inference time. While models like DeepSeek-R1 and o1 have demonstrated that longer thinking leads to better answers, the cost scales linearly with token count — making real-world deployment expensive. This paper proposes a two-stage training pipeline: first a dedicated "semantically-preserved compressor" is trained on annotated mathematical reasoning data, then the base LLM is fine-tuned on compressed CoT pairs using a mixed-ratio SFT regime followed by a novel RL algorithm called CHRPO (Constrained and Hierarchical Ratio Policy Optimization).

Why It Matters

The key innovation here is achieving 73%+ token reduction on MATH-500 with Qwen3-1.7B while actually improving accuracy by 0.6% — a result that sounds paradoxical but reflects a real phenomenon: much of the token budget in standard CoT traces is redundant self-repetition, false starts, and verbose restatements that don't contribute to correct answers. By training the model to internalize a compressed reasoning style, Extra-CoT essentially teaches the model to "think tighter." This is a fundamentally different approach from post-hoc pruning or summarization, since the compression is baked into the generation process itself via RL reward shaping.

Implications for Production AI

The inference efficiency angle here has direct commercial implications. If reasoning models can maintain answer quality at 73% fewer tokens, that translates to roughly 3x throughput improvement and proportional cost reduction — critical for any API provider serving reasoning-heavy workloads at scale. This feeds directly into the cost-collapse narrative currently reshaping the AI industry. Furthermore, CHRPO's hierarchical reward structure — which explicitly penalizes verbosity under tighter budgets — is a clean formulation that could generalize beyond mathematical reasoning to code generation, multi-step planning, and scientific Q&A.

Caveats and Open Questions

The benchmarks are limited to mathematical reasoning tasks (MATH-500 and two others), which may not generalize to domains where chain-of-thought diversity and exploration are more critical — e.g. open-ended scientific reasoning or adversarial tasks. The compressor model itself adds a training complexity that smaller labs may struggle to replicate. And the paper doesn't address whether compressed CoT affects interpretability or auditability — shorter traces may be harder for humans to verify. Still, as a step toward making reasoning models practical, Extra-CoT is a meaningful contribution to the efficiency frontier.

Key Findings
1
73%+ token reduction on MATH-500 with Qwen3-1.7B while improving accuracy by 0.6%
2
CHRPO (Constrained and Hierarchical Ratio Policy Optimization) RL algorithm explicitly incentivizes correct answers under tight token budgets
3
Mixed-ratio SFT training teaches the model to adapt compression level dynamically rather than targeting a single fixed budget
Impact Score
8/10
Expert Commentary
Extra-CoT addresses a real bottleneck in reasoning model deployment — verbose CoT traces that inflate cost without improving accuracy. The 73% compression result is striking and suggests current reasoning traces contain significant redundancy that RL-based training can systematically eliminate.
D
Dong Tran
AI Researcher. Deep-dive analysis of frontier AI research papers.