mlsanityData to 7 Oct 2026

SelfBudgeter: Adaptive Token Allocation for Efficient LLM Reasoning

arXiv 2505.11274 ↗First posted 16 May 2025cs.AI cs.CLcited by 47

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

Abstract (arXiv)

Recently, large reasoning models demonstrate exceptional performance on various tasks. However, reasoning models always consume excessive tokens even for simple queries, leading to resource waste and prolonged user latency. To address this challenge, we propose SelfBudgeter - a self-adaptive reasoning strategy for efficient and controllable reasoning. Specifically, we first train the model to self-estimate the required reasoning budget based on the query. We then introduce budget-guided GRPO for reinforcement learning, which effectively maintains accuracy while reducing output length. Experimental results demonstrate that SelfBudgeter dynamically allocates budgets according to problem complexity, achieving an average response length compression of 61% on math reasoning tasks while maintaining accuracy. Furthermore, SelfBudgeter allows users to see how long generation will take and decide whether to continue or stop. Additionally, users can directly control the reasoning length by setting token budgets upfront.

Key facts. Cited by 47 papers in the map (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 35 of its 41 references that are in the map. First posted on arXiv 16 May 2025.

Builds on

35 of 41 references.

Closest ancestors · 19

Same direction, newest first

  1. 2025Scalable Chain of Thoughts via Elastic Reasoning8 May
  2. 2025Dynamic Early Exit in Reasoning Models22 Apr
  3. 2025SpecReason: Fast and Accurate Inference-Time Compute via Speculative Reasoning10 Apr
  4. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models20 Mar
  5. 2025L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning6 Mar
  6. 2025How Well do LLMs Compress Their Own Chain-of-Thought? A Token Complexity Approach3 Mar
  7. 2025Self-Training Elicits Concise Reasoning in Large Language Models27 Feb
  8. 2025Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning25 Feb
  9. 2025Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities?17 Feb
  10. 2025TokenSkip: Controllable Chain-of-Thought Compression in LLMs17 Feb
  11. 2025CoT-Valve: Length-Compressible Chain-of-Thought Tuning13 Feb
  12. 2025Demystifying Long Chain-of-Thought Reasoning in LLMs5 Feb
7 more
  1. 2025s1: Simple test-time scaling31 Jan
  2. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning22 Jan
  3. 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs30 Dec
  4. 2024Token-Budget-Aware LLM Reasoning24 Dec
  5. 2024C3oT: Generating Shorter Chain-of-Thought without Compromising Effectiveness16 Dec
  6. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug
  7. 2024Concise Thoughts: Impact of Output Length on LLM Reasoning and Cost29 Jul

Other references · 16

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  4. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  5. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,748
  6. 2024RouteLLM: Learning to Route LLMs with Preference Datacited by 633
  7. 2025Process Reinforcement through Implicit Rewardscited by 478
  8. 2025Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Modelscited by 381
  9. 2025Chain of Draft: Thinking Faster by Writing Lesscited by 290
  10. 2025Training Language Models to Reason Efficientlycited by 242
  11. 2025When More is Less: Understanding Chain-of-Thought Length in LLMscited by 182
  12. 2025Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketchingcited by 116
4 more
  1. 2025An Empirical Study on Eliciting and Improving R1-like Reasoning Modelscited by 78
  2. 2024The Benefits of a Concise Chain of Thought on Problem-Solving in Large Language Modelscited by 62
  3. 2024Can Language Models Learn to Skip Steps?cited by 59
  4. 2025Confident or Seek Stronger: Exploring Uncertainty-Based On-device LLM Routing From Benchmarking to Generalizationcited by 29

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 15; 2026: 32.

47 papers.

Sort:

By area:

  1. 2025From Web Search towards Agentic Deep Research: Incentivizing Search with Reasoning Agentscited by 49
  2. 2025Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Modelscited by 36
  3. 2025ParaThinker: Native Parallel Thinking as a New Paradigm to Scale LLM Test-time Computecited by 36
  4. 2026Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgetscited by 28
  5. 2025BudgetThinker: Empowering Budget-aware LLM Reasoning with Control Tokenscited by 27
  6. 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
  7. 2025DRPO: Efficient Reasoning via Decoupled Reward Policy Optimizationcited by 16
  8. 2026Inference-Time Budget Control for LLM Search Agentscited by 13
  9. 2025SABER: Switchable and Balanced Training for Efficient LLM Reasoningcited by 12
  10. 2025From Efficiency to Adaptivity: A Deeper Look at Adaptive Reasoning in Large Language Modelscited by 11
  11. 2026Spend Less, Reason Better: Budget-Aware Value Tree Search for LLM Agentscited by 10
  12. 2025Hierarchical Budget Policy Optimization for Adaptive Reasoningcited by 8
35 more
  1. 2026Adaptive Test-Time Compute Allocation for Reasoning LLMs via Constrained Policy Optimizationcited by 8
  2. 2025Improving Chain-of-Thought Efficiency for Autoregressive Image Generationcited by 7
  3. 2026Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Modelscited by 7
  4. 2026DART: Draft-Agreement Routing for Training-Free Adaptive Thinking Budgets in Hybrid Reasoning Modelscited by 7
  5. 2026Retrospective Progress-Aware Self-Refinement for LLM Agent Trainingcited by 6
  6. 2025Thinking-Free Policy Initialization Makes Distilled Reasoning Models More Effective and Efficient Reasonerscited by 5
  7. 2026FROST: Filtering Reasoning Outliers with Attention for Efficient Reasoningcited by 5
  8. 2026Reward Modeling for Reinforcement Learning-Based LLM Reasoning: Design, Challenges, and Evaluationcited by 5
  9. 2026TRIAGE: Evaluating Prospective Metacognitive Control in LLMs under Resource Constraintscited by 5
  10. 2026AVIS: Adaptive Test-Time Scaling for Vision-Language Modelscited by 4
  11. 2025Concise Reasoning in the Lens of Lagrangian Optimizationcited by 3
  12. 2026Contrastive On-Policy Distillationcited by 3
  13. 2025Adaptive Dual Reasoner: Large Reasoning Models Can Think Efficiently by Hybrid Reasoningcited by 2
  14. 2026SeqRoute: Global Budget-Aware Sequential LLM Routing via Offline Reinforcement Learningcited by 2
  15. 2026The Shadow Price of Reasoning: Economic Perspective on Optimal Budget Allocation for LLMscited by 2
  16. 2026A State-Transition Framework for Efficient LLM Reasoningcited by 1
  17. 2026Structured Prompt Language: Declarative Context Management for LLMscited by 1
  18. 2026Reasoning as Compression: Unifying Budget Forcing via the Conditional Information Bottleneckcited by 1
  19. 2026Can your AI agent be cheaper? Investigating the effects of task specifications on token spend in agentic coding taskscited by 1
  20. 2025Fluidity Index: Next-Generation Super-intelligence Benchmarkscited by 0
  21. 2025ProofSketch: Efficient Verified Reasoning for Large Language Modelscited by 0
  22. 2026Apriel-Reasoner: RL Post-Training for General-Purpose and Efficient Reasoningcited by 0
  23. 2026Batched Contextual Reinforcement: A Task-Scaling Law for Efficient Reasoningcited by 0
  24. 2026Avoiding Overthinking and Underthinking: Curriculum-Aware Budget Scheduling for LLMscited by 0
  25. 2026The Coupling Tax: How Shared Token Budgets Undermine Visible Chain-of-Thought Under Fixed Output Limitscited by 0
  26. 2026LEAD: Length-Efficient Adaptive and Dynamic Reasoning for Large Language Modelscited by 0
  27. 2026CLORE: Content-Level Optimization for Reasoning Efficiencycited by 0
  28. 2026Uncertainty-Aware Budget Allocation for Adaptive Test-Time Reasoningcited by 0
  29. 2026RecurGuard: Runtime Monitoring for Reasoning-Token Consumption Attackscited by 0
  30. 2026Budget-Aware LLM Discovery via Cost-Calibrated Frontier Utilitycited by 0
  31. 2026R3-Bench: LLMs Struggle with Resource-Rational Reasoning under Shared Budgetscited by 0
  32. 2026REChart: Reasoning-Efficient Chart Editing with Large Reasoning Modelscited by 0
  33. 2026AdaThinking-E: One-Token Entropy Regulation for Adaptive Thinkingcited by 0
  34. 2026DRT: Dense Reasoning Trace for Efficient and Grounded Multimodal Reasoningcited by 0
  35. 2026Analyzing and Mitigating Cost-Inefficient Behaviors in Coding Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Reasoning in Token Economies: Budget-Aware Evaluation of LLM Reasoning Strategies
  2. 2025In-Token Rationality Optimization: Towards Accurate and Concise LLM Reasoning via Self-Feedback
  3. 2025Reasoning on a Budget: A Survey of Adaptive and Controllable Test-Time Compute in LLMs
  4. 2025Leash: Adaptive Length Penalty and Reward Shaping for Efficient Large Reasoning Model
  5. 2025Token-Efficient RL for LLM Reasoning
  6. 2025Genius: A Generalizable and Purely Unsupervised Self-Training Framework For Advanced Reasoning
  7. 2025Boosting Accuracy and Efficiency of Budget Forcing in LLMs via Reinforcement Learning for Mathematical Reasoning
  8. 2025Boosting LLM Reasoning via Spontaneous Self-Correction