推理模型论文趋势 2023–2026

链式推理、测试时计算与显式推理能力

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 3,512 篇被归入该主题的论文。

主要结论

每年论文数

各会议论文数

按年份和会议统计

推理模型论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
20232715710933470
2024121846343311
202550020315970932
20263616525822041,799

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

全部推理模型论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 3,512 篇，第 20 / 71 页

  1. ICML 2026Contrastive Reasoning Alignment: Reinforcement Learning from Hidden Representations
  2. ICML 2026ConvexBench: Can LLMs Recognize Convex Functions?
  3. ICML 2026CoRe: Collaborative Reasoning via Cross Teaching
  4. ICML 2026Correcting in Hindsight: Editing Past Key-Value States for Robust LLM Reasoning
  5. ICML 2026CoT is Not the Chain of Truth: An Empirical Internal Analysis of Reasoning LLMs for Fake News Generation
  6. ICML 2026Counterfactual Residual Data Augmentation for Regression
  7. ICML 2026Coupled Variational Reinforcement Learning for Language Model General Reasoning
  8. ICML 2026CPMöbius: Iterative Coach–Player Reasoning for Data-Free Reinforcement Learning
  9. ICML 2026Critique-Guided Distillation for Robust Reasoning via Refinement
  10. ICML 2026CyberJurors: A Multi-Agent Simulation Task for E-Commerce Disputes Verdict
  11. ICML 2026D-CORE: Incentivizing Task Decomposition in Large Reasoning Models for Complex Tool Use
  12. ICML 2026D-FUSEr: Diverse Failure, Unified Success via Error-Distribution Shaping in LLM Reasoning
  13. ICML 2026D²Evo: Dual Difficulty-Aware Self-Evolution for Data-Efficient Reinforcement Learning
  14. ICML 2026Data Provenance Auditing of Fine-Tuned Large Language Models with a Text-Preserving Technique
  15. ICML 2026DC-W2S: Dual-Consensus Weak-to-Strong Training for Reliable Process Reward Modeling in Biological Reasoning
  16. ICML 2026DecepChain: Inducing Deceptive Reasoning in Large Language Models
  17. ICML 2026Decompose and Recompose: Reasoning New Skills from Existing Abilities for Cross-Task Robotic Manipulation
  18. ICML 2026Decomposed On-Policy Distillation for Vision-Language Reasoning: Steering Gradients for Visual Grounding
  19. ICML 2026Deep Scientific Reasoning under Physical Constraints: Structure-Aware Spectrum Prediction
  20. ICML 2026Demystifying Scientific Problem-Solving in LLMs by Probing Knowledge and Reasoning
  21. ICML 2026DenseSteer: Steering Small Language Models towards Dense Math Reasoning
  22. ICML 2026Depth-Breadth Synergy in RLVR: Unlocking LLM Reasoning Gains with Adaptive Exploration
  23. ICML 2026Differential Smoothing Mitigates Sharpening and Improves LLM Reasoning
  24. ICML 2026Directly Optimizing Natural Language Explanations for Behavioral Faithfulness: Simulatability and Recoverability
  25. ICML 2026Discrete Adjoint Schrödinger Bridge Sampler
  26. ICML 2026Discrete Diffusion with Physical Mass Constraints for \emph{De Novo} Peptide Sequencing
  27. ICML 2026Discriminative Visual Process Rewards for Scaling Thinking at Test-Time with Images
  28. ICML 2026DisPPO: Quantile-Based Distributional Reinforcement Learning for Large Language Models
  29. ICML 2026Distilling Neuro-Symbolic Programs into 3D Multi-modal LLMs
  30. ICML 2026Distilling Task-Level Coordination Policies for Generalizable Multi-Agent Cooperation
  31. ICML 2026Diversity Matters: Revisiting Test-Time Compute in Vision-Language Models
  32. ICML 2026Diversity Over Frequency: Rethinking Tool Use in Visual Chain-of-Thought Agents
  33. ICML 2026Do LLMs Signal When They’re Right? Evidence from Neuron Agreement
  34. ICML 2026Do Sparse Autoencoders Identify Reasoning Features in Language Models?
  35. ICML 2026Do Vision and Text Cues Exhibit Evidential Coupling? UFO: A Benchmark for Compositional Multimodal Reasoning in Unified Models
  36. ICML 2026DocHop: Benchmarking Out-of-domain Multi-hop Reasoning in Information-Dense Documents
  37. ICML 2026DocVAL: Validated Chain-of-Thought Distillation for Grounded Document VQA
  38. ICML 2026Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning
  39. ICML 2026Does Reasoning Improve Seeing? Understanding When Vision-Language Models Benefit from Thinking
  40. ICML 2026Does Your Reasoning Model Implicitly Know When to Stop Thinking?
  41. ICML 2026Don't Force the Fit: Bounded Log-Likelihood Loss for Enhanced Reasoning in Large Language Models
  42. ICML 2026Don't Ignore the Tail: Decoupled Distillation Produces Top Maths Students on an Academic Budget
  43. ICML 2026DREAM-R: Multimodal Speculative Reasoning with RL-Based Refined Drafting, Precise Verification, and Fully Parallel Execution
  44. ICML 2026DRIVE: Best Data Scheduling Practices for Reinforcement Learning with Verifiable Reward in Competitive Code Generation
  45. ICML 2026DTS: Enhancing Large Reasoning Models via Decoding Tree Sketching
  46. ICML 2026Dual-Latent Memory Routing for Vision-Language Reasoning
  47. ICML 2026DyCon: Dynamic Reasoning Control via Evolving Difficulty Modeling
  48. ICML 2026Dynamic Thinking-Token Selection for Efficient Reasoning in Large Reasoning Models
  49. ICML 2026Dynamics Reveals Structure: Challenging the Linear Propagation Assumption
  50. ICML 2026Dynamics Within Latent Chain-of-Thought: An Empirical Study of Causal Structure

在 Claude、Cursor 里直接问这个趋势

接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。

可以直接复制这句话问 AI

用 paperlist 查推理模型在 NeurIPS 2023–2026 的论文趋势，哪一年涨得最快？

AI 会调用的工具：get_topic_stats {"topic": "reasoning-models"}

两分钟接入 Paperlist MCP（Claude Code、Cursor、Codex、ChatGPT）

常见问题

2023–2026 年四大会议一共有多少篇推理模型论文？
共 3,512 篇，按年份分别是 2023 年 470 篇、2024 年 311 篇、2025 年 932 篇、2026 年 1,799 篇。
NeurIPS 2026 有多少篇推理模型论文？
NeurIPS 2026 有 361 篇推理模型论文，上一届是 500 篇。
ICLR 2026 有多少篇推理模型论文？
ICLR 2026 有 652 篇推理模型论文，上一届是 203 篇。
ICML 2026 有多少篇推理模型论文？
ICML 2026 有 582 篇推理模型论文，上一届是 159 篇。
CVPR 2026 有多少篇推理模型论文？
CVPR 2026 有 204 篇推理模型论文，上一届是 70 篇。
哪个会议的推理模型论文最多？
NeurIPS，2023–2026 年共 1,253 篇。

「前沿系统」下的其他主题