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 1,799 篇，第 32 / 36 页

  1. NeurIPS 2026ExpLang: Improved Exploration and Exploitation in LLM Reasoning with On-Policy Thinking Language Selection
  2. NeurIPS 2026Explicit Geometric Chain-of-Thought for Vision-Language-Action in Autonomous Driving
  3. NeurIPS 2026Extracting Search Trees from LLM Reasoning Traces Reveals Myopic Planning
  4. NeurIPS 2026Facts Don't Speak Louder than Words: The Behavioral Essence of Long-CoT Distillation
  5. NeurIPS 2026Failure Profiling and Reachable Trajectory Selection for Reasoning Distillation
  6. NeurIPS 2026Fill the GAP: A Granular Alignment Paradigm for Visual Reasoning in Multimodal Large Language Models
  7. NeurIPS 2026Fixed Universal Transformers
  8. NeurIPS 2026Fixed-Point Reasoning: Stable and Adaptive Deep Looped Models
  9. NeurIPS 2026Forecasting Downstream Performance of LLMs With Proxy Metrics
  10. NeurIPS 2026From Cortical Synchronous Rhythm to Brain Inspired Learning Mechanism: An Oscillatory Spiking Neural Network with Time-Delayed Coordination
  11. NeurIPS 2026From Detection to Understanding — A Multi-Task Dataset for Traffic Anomaly Reasoning
  12. NeurIPS 2026From Document Layout to Causal Topology: An End-to-End Architecture for Temporal Causal Reasoning
  13. NeurIPS 2026From Outcome to Representation: Tracing Reasoning Mechanisms through Integrated Policy Gradient
  14. NeurIPS 2026From Perception to Punchline: Empowering VLM with the Art of In-the-wild Memes
  15. NeurIPS 2026From Solver Trajectories to Teaching Trajectories: Cognition-Aligned Reasoning Distillation
  16. NeurIPS 2026GeLVR: Geometry-Consistent Latent Visual Reasoning in Multimodal LLMs
  17. NeurIPS 2026Generalizing Test-time Compute-optimal Scaling as an Optimizable Graph
  18. NeurIPS 2026Geometric Factual Recall in Transformers
  19. NeurIPS 2026Geometric Latent Reasoning Induces Shorter Generations in LLMs
  20. NeurIPS 2026Geometric Prompt-Trajectory Planning for Test-Time Scaling
  21. NeurIPS 2026Geometric Signatures of Reasoning: A Spectral Perspective on Task Hardness
  22. NeurIPS 2026GeoSym127K: Scalable Symbolically-verifiable Synthesis for Multimodal Geometric Reasoning
  23. NeurIPS 2026Global Importance Estimation for KV Cache Eviction
  24. NeurIPS 2026GOLD PANNING: Strategic Context Shuffling for Needle-in-Haystack Reasoning
  25. NeurIPS 2026GPA: Generative Population Annealing for Test-Time Sequence Design with Pretrained Generative Models
  26. NeurIPS 2026GridProbe: Posterior-Probing for Adaptive Test-Time Compute in Long-Video VLMs
  27. NeurIPS 2026GUI-Libra: Data-Efficient Post-Training for Reliable Reasoning-and-Acting in Native GUI Agents
  28. NeurIPS 2026Hawkeye: Hardware-Aware GPU Kernel Optimization with Minimal Supervision
  29. NeurIPS 2026Hear, Localize, and Reason: Spatially Aware Scene Understanding for Audio-visual LLMs
  30. NeurIPS 2026Hint Tuning: Less Data Makes Better Reasoners
  31. NeurIPS 2026Hitting a Moving Target: Test-Time Adaptation for AI Text Detection under Continual Distribution Shift
  32. NeurIPS 2026HopChain: Multi-Hop Data Synthesis for Generalizable Vision-Language Reasoning
  33. NeurIPS 2026Hyperbolic Language Models: From Zipf to Compute-Optimal Scaling
  34. NeurIPS 2026Improving Audit Realism with Inference-Time Compute and Deployment Scaffolds
  35. NeurIPS 2026Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
  36. NeurIPS 2026InfoSFT: Learn More and Forget Less with Information-Aware Token Weighting
  37. NeurIPS 2026INFUSER: Influence-Guided Self-Evolution Improves Reasoning
  38. NeurIPS 2026Intent Factored Generation: Unleashing the Diversity in Your Language Model
  39. NeurIPS 2026Interleaved Latent Thinking and Adaptive Termination for Efficient Reasoning LLMs
  40. NeurIPS 2026Iterative Nonlinear Computation Underlying Abstract Reasoning
  41. NeurIPS 2026Joint Consistency: A Unified Test-Time Aggregation Framework via Energy Minimization
  42. NeurIPS 2026Language Models Need Sleep
  43. NeurIPS 2026Large-Scale Pretraining unlocks Few-Shot Prediction for Relational Data
  44. NeurIPS 2026Latent Reasoning in Continuous Space for Unified Multimodal Models
  45. NeurIPS 2026Latent-space Attacks for Refusal Evasion in Language Models
  46. NeurIPS 2026LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning
  47. NeurIPS 2026LeAct: Learning to Reason from Expert Actions
  48. NeurIPS 2026LEAD: Length-Efficient Adaptive and Dynamic Reasoning for Large Language Models
  49. NeurIPS 2026Learning from Trials and Errors: Reflective Test-Time Planning for Embodied LLMs
  50. NeurIPS 2026Learning Robust Reasoning through Guided Adversarial Self-Play

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