mlsanityData to 14 Oct 2026

Your Models Have Thought Enough: Training Large Reasoning Models to Stop Overthinking

arXiv 2509.23392 ↗First posted 27 Sep 2025cs.AI cs.CLcited by 10

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Large Reasoning Models (LRMs) have achieved impressive performance on challenging tasks, yet their deep reasoning often incurs substantial computational costs. To achieve efficient reasoning, existing reinforcement learning methods still struggle to construct short reasoning path during the rollout stage, limiting effective learning. Inspired by Evidence Accumulation Models, we find that LRMs have accumulated sufficient information early in reasoning, making further reasoning steps redundant. Based on this insight, we propose Just-Enough Thinking (JET), which trains models to proactively terminate unnecessary reasoning. JET performs trajectory truncation during rollout to expose the model to short, distributionally consistent reasoning paths. Besides, it uses a quality-controlled length reward to better encourage concise reasoning while maintaining correctness. Extensive experiments demonstrate that JET significantly improves reasoning efficiency without sacrificing accuracy. Especially, DeepSeek-Distill-Qwen-1.5B achieves a 4.6% accuracy gain while reducing output length by 46.3% on the Olympiad benchmark. Our code is available in the GitHub.

Key facts. Cited by 10 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 42 of its 50 references that are in the map. First posted on arXiv 27 Sep 2025.

Builds on

42 of 50 references.

Closest ancestors · 18

Same direction, newest first

  1. 2025Mitigating Strategy-Selection Bias in Reasoning for More Effective Test-Time Scaling22 Sep
  2. 2025How Far Are We from Optimal Reasoning Efficiency?8 Jun
  3. 2025AdaCtrl: Towards Adaptive and Controllable Reasoning via Difficulty-Aware Budgeting24 May
  4. 2025AdaptThink: Reasoning Models Can Learn When to Think19 May
  5. 2025AdaCoT: Pareto-Optimal Adaptive Chain-of-Thought Triggering via Reinforcement Learning17 May
  6. 2025An Empirical Study of LLM Reasoning Ability Under Strict Output Length Constraint19 Apr
  7. 2025Efficient Reasoning Models: A Survey15 Apr
  8. 2025ThinkPrune: Pruning Long Chain-of-Thought of LLMs via Reinforcement Learning2 Apr
  9. 2025A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyond27 Mar
  10. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models20 Mar
  11. 2025L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning6 Mar
  12. 2025DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models6 Mar
6 more
  1. 2025CoT-Valve: Length-Compressible Chain-of-Thought Tuning13 Feb
  2. 2025s1: Simple test-time scaling31 Jan
  3. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning22 Jan
  4. 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs30 Dec
  5. 2024Imitate, Explore, and Self-Improve: A Reproduction Report on Slow-thinking Reasoning Systems12 Dec
  6. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug

Other references · 24

  1. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  2. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  4. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
  5. 2023Qwen Technical Reportcited by 3,344
  6. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
  7. 2025Understanding R1-Zero-Like Training: A Critical Perspectivecited by 1,345
  8. 2024OpenAI o1 System Cardcited by 1,244
  9. 2024OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problemscited by 1,136
  10. 2024Phi-4 Technical Reportcited by 711
  11. 2025SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-trainingcited by 672
  12. 2024A Survey on Large Language Models for Code Generationcited by 658
12 more
  1. 2025Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Modelcited by 461
  2. 2025Kimi-VL Technical Reportcited by 388
  3. 2025From System 1 to System 2: A Survey of Reasoning Large Language Modelscited by 320
  4. 2025Training Language Models to Reason Efficientlycited by 247
  5. 2025REINFORCE++: Stabilizing Critic-Free Policy Optimization with Global Advantage Normalizationcited by 241
  6. 2025Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMscited by 172
  7. 2025Optimizing Test-Time Compute via Meta Reinforcement Fine-Tuningcited by 123
  8. 2025S-GRPO: Early Exit via Reinforcement Learning in Reasoning Modelscited by 87
  9. 2025Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognitioncited by 62
  10. 2025Blending Supervised and Reinforcement Fine-Tuning with Prefix Samplingcited by 60
  11. 2025ARM: Adaptive Reasoning Modelcited by 39
  12. 2025Ada-R1: Hybrid-CoT via Bi-Level Adaptive Reasoning Optimizationcited by 33

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 10.

10 papers.

  1. 2026Retrieval--Reasoning Processes for Multi-hop Question Answering: A Four-Axis Design Framework and Empirical Trendscited by 8
  2. 2026Mitigating Overthinking in Large Reasoning Models via Difficulty-aware Reinforcement Learningcited by 6
  3. 2026From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AIcited by 6
  4. 2026Mitigating Factual Hallucination in Large Reasoning Models via Mixed-Mode Advantage Regularizationcited by 3
  5. 2026TRiMS: Real-Time Tracking of Minimal Sufficient Length for Efficient Reasoning via RLcited by 2
  6. 2026ExpThink: Experience-Guided Reinforcement Learning for Adaptive Chain-of-Thought Compressioncited by 2
  7. 2026ADaPT: Token-Level Decoupling for Efficient Large Reasoning Modelscited by 0
  8. 2026Do small language models know what they don't know?cited by 0
  9. 2026Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiencycited by 0
  10. 2026Quality Determines Direction, Length Shapes Magnitude: Length Control for Open-Ended Reinforcement Learningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Models
  2. 2025Large Reasoning Models are not thinking straight: on the unreliability of thinking trajectories
  3. 2025Gold-Switch: Training-Free Superposition of Slow- and Fast- Thinking LLMs
  4. 2025Don't Think Longer, Think Wisely: Optimizing Thinking Dynamics for Large Reasoning Models
  5. 2025FlashThink: An Early Exit Method For Efficient Reasoning
  6. 2025Exploring and Exploiting the Inherent Efficiency within Large Reasoning Models for Self-Guided Efficiency Enhancement
  7. 2025Learning to Reason Efficiently with Discounted Reinforcement Learning
  8. 2025Mitigating Overthinking through Reasoning Shaping