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 361 篇，第 7 / 8 页

  1. NeurIPS 2026Structured Sparse Memory for Recurrent Reasoning
  2. NeurIPS 2026SWE-Chain: Benchmarking Coding Agents on Chained Release-Level Package Upgrades
  3. NeurIPS 2026Taming CoT Obfuscation in VLMs: From Mechanistic Evidence to Activation-Level Enforcement
  4. NeurIPS 2026TaskGround: Structured Executable Task Inference for Full-Scene Household Reasoning
  5. NeurIPS 2026Tatemae: Detecting Alignment Faking via Tool Selection in LLMs
  6. NeurIPS 2026Teach-to-Reason: Competition-Guided Reasoning with a Self-Improving Teacher
  7. NeurIPS 2026Teaching Large Language Models When Not to Know: Learning Temporal Critique for Ex-Ante Reasoning
  8. NeurIPS 2026Teaching VLMs What to Say, Not How to Reason: Rethinking Counterfactual Reasoning in Autonomous Driving
  9. NeurIPS 2026Temporal Backtracking Search for Test-time Generative Video Reasoning
  10. NeurIPS 2026TERMINATOR: Learning Optimal Exit Points for Early Stopping in Chain-of-Thought Reasoning
  11. NeurIPS 2026Test Time Search Requires Training for Diversity
  12. NeurIPS 2026Test time training enhances in-context learning of nonlinear functions
  13. NeurIPS 2026Test-Time Compute Games
  14. NeurIPS 2026Test-Time Graph Anomaly Detection via Shifted Augmentation with Dynamic Objective Scheduling
  15. NeurIPS 2026Test-Time Prompt-Agnostic Decomposition
  16. NeurIPS 2026Test-Time Scaling in the Wild: Why Exploitation, Not Exploration, Is the Bottleneck
  17. NeurIPS 2026The Capability Frontier of GRPO
  18. NeurIPS 2026The Communication Bottleneck: A Round-Trip Study of Compositional Serialization in Language Models
  19. NeurIPS 2026Thinking with Images as Continuous Policy: Numerical Visual Chain-of-Thought
  20. NeurIPS 2026Thinking with Imagination: Agentic Visual Spatial Reasoning with World Simulators
  21. NeurIPS 2026Thinking with Visual Primitives
  22. NeurIPS 2026ThinkJEPA: Empowering Latent World Models with Large Vision-Language Reasoning Model
  23. NeurIPS 2026ThinkSafe: Self-Generated Safety Alignment for Reasoning Models
  24. NeurIPS 2026TIDES: Implicit Time-Awareness in Selective State Space Models
  25. NeurIPS 2026Tio: Language Models with Parallel Streams of Thoughts, Inputs and Outputs
  26. NeurIPS 2026To Think or Not to Think: Pre-Decisional Reasoning Budgets for Referring Audio-Visual Segmentation
  27. NeurIPS 2026TokenSwap: Benchmarking and Reducing the Modality Gap in Multimodal LLMs
  28. NeurIPS 2026Tool Verification for Test-Time Reinforcement Learning
  29. NeurIPS 2026Topological Periodicity Test (TopPT) via Confidence Bound of Time-Delay Embeddings
  30. NeurIPS 2026Topology-Aware Optimal Transport for Source-Free Test-Time Adaptation in Anomaly Segmentation
  31. NeurIPS 2026Toward Cognitive Supersensing in Multimodal Large Language Models
  32. NeurIPS 2026Towards Error-Free EHRs: Reasoning-Intensive Consistency Verification Between Clinical Notes and Structured Tables in Electronic Health Records
  33. NeurIPS 2026Towards Generalizable Data Valuation: Learning an End-to-end Deep Model to Compute Shapley Values
  34. NeurIPS 2026Towards Principled Efficient Rollout Allocation in Test-Time Training
  35. NeurIPS 2026Train at the Moving Edge: Rollout-Efficient RL for Large Reasoning Models
  36. NeurIPS 2026Training on Documents About Monitoring Leads to CoT Obfuscation
  37. NeurIPS 2026Training Quality Determines Efficiency Boundaries in Test-Time Reasoning
  38. NeurIPS 2026Training with Harnesses: On-Policy Harness Self-Distillation for Complex Reasoning
  39. NeurIPS 2026TTB: Test-time MLP Baking for Efficient Rendering of Decoder-only View Synthesis Models
  40. NeurIPS 2026Understanding the Out-of-Distribution Generalization of Chain-of-Thought Reasoning in LLMs
  41. NeurIPS 2026Uniform-in-Time Weak Propagation of Chaos in Shallow Neural Networks
  42. NeurIPS 2026Unifying Reasoning and Planning through Energy Minimization
  43. NeurIPS 2026UniPath: Adaptive Coordination of Understanding and Generation for Unified Multimodal Reasoning
  44. NeurIPS 2026Universal and Efficient Computation with 2D Attention
  45. NeurIPS 2026UniVLR: Unifying Text and Vision in Visual Latent Reasoning for Multimodal LLMs
  46. NeurIPS 2026Value-Aware Stochastic KV Cache Eviction for Reasoning Models
  47. NeurIPS 2026Vision to Geometry: 3D Spatial Memory for Sequential Embodied MLLM Reasoning and Exploration
  48. NeurIPS 2026Visual Enhanced Depth Scaling for Multimodal Latent Reasoning
  49. NeurIPS 2026Voronoi-Markov chain and spatial entropy for point pattern analysis
  50. NeurIPS 2026Walking Through 3D Spaces: Spatial Routing for Referring 3D Gaussian Splatting Segmentation

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