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 500 篇，第 10 / 10 页

  1. NeurIPS 2025TopoPoint: Enhance Topology Reasoning via Endpoint Detection in Autonomous Driving
  2. NeurIPS 2025Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning
  3. NeurIPS 2025Tracking and Understanding Object Transformations
  4. NeurIPS 2025Training a Scientific Reasoning Model for Chemistry
  5. NeurIPS 2025Training Language Models to Reason Efficiently
  6. NeurIPS 2025Trajectory Balance with Asynchrony: Decoupling Exploration and Learning for Fast, Scalable LLM Post-Training
  7. NeurIPS 2025Transformers Provably Learn Chain-of-Thought Reasoning with Length Generalization
  8. NeurIPS 2025Tropical Attention: Neural Algorithmic Reasoning for Combinatorial Algorithms
  9. NeurIPS 2025Trust, But Verify: A Self-Verification Approach to Reinforcement Learning with Verifiable Rewards
  10. NeurIPS 2025TRUST: Test-Time Refinement using Uncertainty-Guided SSM Traverses
  11. NeurIPS 2025Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Training
  12. NeurIPS 2025UFT: Unifying Supervised and Reinforcement Fine-Tuning
  13. NeurIPS 2025Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inference
  14. NeurIPS 2025Uni-MuMER: Unified Multi-Task Fine-Tuning of Vision-Language Model for Handwritten Mathematical Expression Recognition
  15. NeurIPS 2025Unified Multimodal Chain-of-Thought Reward Model through Reinforcement Fine-Tuning
  16. NeurIPS 2025UniGen: Enhanced Training & Test-Time Strategies for Unified Multimodal Understanding and Generation
  17. NeurIPS 2025UniPixel: Unified Object Referring and Segmentation for Pixel-Level Visual Reasoning
  18. NeurIPS 2025UniTransfer: Video Concept Transfer via Progressive Spatio-Temporal Decomposition
  19. NeurIPS 2025Unlabeled Data Can Provably Enhance In-Context Learning of Transformers
  20. NeurIPS 2025Unlocking Multimodal Mathematical Reasoning via Process Reward Model
  21. NeurIPS 2025Unveiling Chain of Step Reasoning for Vision-Language Models with Fine-grained Rewards
  22. NeurIPS 2025Unveiling the Compositional Ability Gap in Vision-Language Reasoning Model
  23. NeurIPS 2025Vad-R1: Towards Video Anomaly Reasoning via Perception-to-Cognition Chain-of-Thought
  24. NeurIPS 2025Value-Guided Search for Efficient Chain-of-Thought Reasoning
  25. NeurIPS 2025VeriThinker: Learning to Verify Makes Reasoning Model Efficient
  26. NeurIPS 2025VeriThoughts: Enabling Automated Verilog Code Generation using Reasoning and Formal Verification
  27. NeurIPS 2025ViCrit: A Verifiable Reinforcement Learning Proxy Task for Visual Perception in VLMs
  28. NeurIPS 2025Video-R1: Reinforcing Video Reasoning in MLLMs
  29. NeurIPS 2025VideoChat-R1.5: Visual Test-Time Scaling to Reinforce Multimodal Reasoning by Iterative Perception
  30. NeurIPS 2025VideoRFT: Incentivizing Video Reasoning Capability in MLLMs via Reinforced Fine-Tuning
  31. NeurIPS 2025Vinci: Deep Thinking in Text-to-Image Generation using Unified Model with Reinforcement Learning
  32. NeurIPS 2025VisDiff: SDF-Guided Polygon Generation for Visibility Reconstruction, Characterization and Recognition
  33. NeurIPS 2025Visual Structures Help Visual Reasoning: Addressing the Binding Problem in LVLMs
  34. NeurIPS 2025Visual Thoughts: A Unified Perspective of Understanding Multimodal Chain-of-Thought
  35. NeurIPS 2025VisualQuality-R1: Reasoning-Induced Image Quality Assessment via Reinforcement Learning to Rank
  36. NeurIPS 2025VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning
  37. NeurIPS 2025VLM-R³: Region Recognition, Reasoning, and Refinement for Enhanced Multimodal Chain-of-Thought
  38. NeurIPS 2025VLMs have Tunnel Vision: Evaluating Nonlocal Visual Reasoning in Leading VLMs
  39. NeurIPS 2025VPO: Reasoning Preferences Optimization Based on $\mathcal{V}$-Usable Information
  40. NeurIPS 2025Walking the Tightrope: Autonomous Disentangling Beneficial and Detrimental Drifts in Non-Stationary Custom-Tuning
  41. NeurIPS 2025Weaver: Shrinking the Generation-Verification Gap by Scaling Compute for Verification
  42. NeurIPS 2025WebThinker: Empowering Large Reasoning Models with Deep Research Capability
  43. NeurIPS 2025What’s in Common? Multimodal Models Hallucinate When Reasoning Across Scenes
  44. NeurIPS 2025When Less Language is More: Language-Reasoning Disentanglement Makes LLMs Better Multilingual Reasoners
  45. NeurIPS 2025When No Paths Lead to Rome: Benchmarking Systematic Neural Relational Reasoning
  46. NeurIPS 2025When Thinking Drifts: Evidential Grounding for Robust Video Reasoning
  47. NeurIPS 2025When Thinking Fails: The Pitfalls of Reasoning for Instruction-Following in LLMs
  48. NeurIPS 2025Which Data Attributes Stimulate Math and Code Reasoning? An Investigation via Influence Functions
  49. NeurIPS 2025Who Reasons in the Large Language Models?
  50. NeurIPS 2025Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree Search

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