When2Think: Learning Difficulty-Aware Length Control for Efficient Hybrid Reasoning Models
Summary
When2Think is a post-training framework that dynamically allocates computation in large reasoning models based on problem difficulty, improving accuracy-efficiency trade-offs on mathematical benchmarks.
View Cached Full Text
Cached at: 09/18/26, 02:58 AM
Paper page - When2Think: Learning Difficulty-Aware Length Control for Efficient Hybrid Reasoning Models
Source: https://huggingface.co/papers/2609.19671
Abstract
LargeReasoningModels(LRMs)achievestrongperformanceoncomplextasksbutexhibitsystematicinefficiency:theyoftenoverthinkeasyproblemsandunderthinkhardones.Existingapproachesbasedonuniformlengthpenaltiesorrigidroutingincuranefficiencytax,tradingreducedcomputationoneasyinstancesforaccuracylossonhardinstances.Weformulateefficientreasoningasaninstance-adaptivecomputationallocationproblemandproposeWhen2Think,apost-trainingframeworkforhybridreasoningthatdynamicallyallocatescomputationbasedonproblemdifficulty.OurmethodintroducesInstance-levelDifficulty-AwareControl(IDAC),areward-shapingmechanismthatleveragespre-computedreferencestatistics(accuracyandtokenusage)toregulatereasoningdepth.Combinedwithverifier-basedrewardsandbatch-wisestandardizedadvantages,IDACenablesstablecritic-freeoptimizationwithoutlearnedrewardmodelsoronlinereference-modelqueries.When2Thinkencouragesdirectansweringoneasyinstanceswhilepreservingextendedreasoningonhardinstances,therebylearningwhentouseSystem1(NoThink)versusSystem2(Think).Experimentsonmathematicalbenchmarksdemonstrateimprovedaccuracy-efficiencytrade-offs:onAIME24,[email protected]%whiletokenusageisreducedby27.9%relativetothebasemodel,andonAIME25,When2Thinkachieves40.0%Pass@3,outperformingcompressionandrouting-onlybaselines.
View arXiv pageView PDFAdd to collection
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2609.19671 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2609.19671 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2609.19671 in a Space README.md to link it from this page.
Collections including this paper1
Similar Articles
LEAD: Length-Efficient Adaptive and Dynamic Reasoning for Large Language Models
LEAD dynamically adapts reasoning efficiency during training by using online calibration of correctness-efficiency trade-offs and adaptive problem-specific length targets, improving mathematical reasoning accuracy and reducing output length.
DyCon: Dynamic Reasoning Control via Evolving Difficulty Modeling
This paper introduces DyCon, a training-free framework that uses step-level embeddings to model evolving task difficulty and dynamically control reasoning depth in Large Reasoning Models, effectively reducing overthinking and improving efficiency without sacrificing accuracy.
Lightning Weave: Improving the Accuracy-Efficiency Frontier of Reasoning Models through Capability Composition
This paper introduces Lightning Weave, a framework that combines separately trained reasoning capabilities into one efficient model using on-policy distillation, enhancing accuracy and reducing token usage in math and code tasks.
A*-Thought-V2: Efficient Latent Reasoning via Geometric Dynamics of LLM
A*-Thought-V2 models chain-of-thought reasoning as hidden-state trajectories, using geometric dynamics to compress non-essential steps into latent tokens, improving accuracy and efficiency in LLM reasoning.
Efficient Reasoning Distillation: Small Video-Language Models via Synthetic CoT and Difficulty-Aware Fine-Tuning
The paper presents a method to distill reasoning into compact video-language models using synthetic chain-of-thought rationales and difficulty-aware fine-tuning, enabling smaller models to outperform larger ones with minimal compute.