Pith. sign in

REVIEW 13 cited by

AdaptThink: Reasoning Models Can Learn When to Think

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2505.13417 v1 pith:74QHZJ7T submitted 2025-05-19 cs.CL cs.AIcs.LG

AdaptThink: Reasoning Models Can Learn When to Think

classification cs.CL cs.AIcs.LG
keywords adaptthinkthinkingreasoningmodelsperformanceefficiencymodelnothinking
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Recently, large reasoning models have achieved impressive performance on various tasks by employing human-like deep thinking. However, the lengthy thinking process substantially increases inference overhead, making efficiency a critical bottleneck. In this work, we first demonstrate that NoThinking, which prompts the reasoning model to skip thinking and directly generate the final solution, is a better choice for relatively simple tasks in terms of both performance and efficiency. Motivated by this, we propose AdaptThink, a novel RL algorithm to teach reasoning models to choose the optimal thinking mode adaptively based on problem difficulty. Specifically, AdaptThink features two core components: (1) a constrained optimization objective that encourages the model to choose NoThinking while maintaining the overall performance; (2) an importance sampling strategy that balances Thinking and NoThinking samples during on-policy training, thereby enabling cold start and allowing the model to explore and exploit both thinking modes throughout the training process. Our experiments indicate that AdaptThink significantly reduces the inference costs while further enhancing performance. Notably, on three math datasets, AdaptThink reduces the average response length of DeepSeek-R1-Distill-Qwen-1.5B by 53% and improves its accuracy by 2.4%, highlighting the promise of adaptive thinking-mode selection for optimizing the balance between reasoning quality and efficiency. Our codes and models are available at https://github.com/THU-KEG/AdaptThink.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 13 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. AdaThinkV: Adaptive Thinking for Token-Efficient Video Reasoning

    cs.CV 2026-08 conditional novelty 7.0

    A video reasoning model learns per question whether to reason aloud or answer directly, improving accuracy by about 3 points over the best adaptive baseline while using about 23% fewer output tokens.

  2. KCSAT-ML: Probing Reasoning Models with Nationwide-Cohort Human Difficulty

    cs.CL 2026-06 unverdicted novelty 7.0

    KCSAT-ML benchmark supplies human error rates for math problems and DRG metric exposes that model accuracy collapses on high-human-error items while test-time scaling shows non-monotonic gains and alignment failures.

  3. An Efficient Streaming Video Understanding Framework with Agentic Control

    cs.CV 2026-05 unverdicted novelty 7.0

    R3-Streaming uses cascaded control with age-aware memory forgetting and TB-GRPO reinforcement learning to reach SOTA scores of 57.92 on OVO-Bench and 76.36 on StreamingBench with 95-96% fewer visual tokens.

  4. From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning

    cs.LG 2026-06 conditional novelty 6.5

    SFT supplies entangled compositional traces of atomic skills and routing modules; RL identifies those modules and enables recombination on novel compositions outside the SFT support.

  5. From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning

    cs.LG 2026-06 unverdicted novelty 6.0

    Introduces a hierarchical latent selection model showing SFT supplies raw module materials in compound traces while RL decomposes them to identify atomic modules and enable recombination for new reasoning configurations.

  6. An Efficient Streaming Video Understanding Framework with Agentic Control

    cs.CV 2026-05 unverdicted novelty 6.0

    R3-Streaming uses cascaded control, age-aware memory forgetting, and TB-GRPO reinforcement learning to reach SOTA scores on streaming video benchmarks while cutting visual token usage by 95-96%.

  7. Length Value Model: Scalable Value Pretraining for Token-Level Length Modeling

    cs.CL 2026-04 unverdicted novelty 6.0

    LenVM models token-level remaining generation length as a bounded discounted value function derived from constant negative per-token rewards, providing a scalable proxy for generation horizon.

  8. Length Value Model: Scalable Value Pretraining for Token-Level Length Modeling

    cs.CL 2026-04 conditional novelty 6.0

    LenVM trains a token-level value head to predict discounted remaining length, enabling length control and efficiency steering on LLMs and VLMs.

  9. HiRO-Nav: Hybrid ReasOning Enables Efficient Embodied Navigation

    cs.AI 2026-04 unverdicted novelty 6.0

    HiRO-Nav adaptively triggers reasoning only on high-entropy actions via a hybrid training pipeline and shows better success-token trade-offs than always-reason or never-reason baselines on the CHORES-S benchmark.

  10. ConPress: Learning Efficient Reasoning from Multi-Question Contextual Pressure

    cs.CL 2026-02 conditional novelty 6.0

    Multi-question prompts elicit shorter chain-of-thought traces, and fine-tuning on those traces transfers the compression to single-question reasoning.

  11. Is Your Model Thinking or Just Stagnating? PUMA: Diagnosing Reasoning Pathology via Phase-Momentum Alignment

    cs.AI 2026-07 conditional novelty 5.0

    PUMA detects when a reasoning model's entropy drop aligns with hidden-state momentum, truncates at that point, and reports improved accuracy-efficiency on 1.5B-32B reasoning models.

  12. VisionPulse: Dynamic Visual Sparsity for Efficient Multimodal Reasoning

    cs.CV 2026-05 unverdicted novelty 5.0

    VisionPulse is a step-wise visual token pruning method for LMMs that retains 5% of tokens per step, shortens reasoning traces by 11.2%, and maintains accuracy.

  13. Reinforcement Learning Meets Large Language Models: A Survey of Advancements and Applications Across the LLM Lifecycle

    cs.CL 2025-09 conditional novelty 3.0

    A survey that maps reinforcement learning methods, datasets, benchmarks, and open-source tools across the full training lifecycle of large language models, focusing on verifiable-reward reasoning.