推理模型论文趋势 2023–2026

主要结论

• 推理模型论文从 2023 年的 470 篇增长到 2026 年的 1,799 篇，三年间扩大到约 3.8 倍。
• NeurIPS 是发推理模型论文最多的会议，2023–2026 年共 1,253 篇，占四个会议总量的 36%。
• 交叉方向里涨得最快的是强化学习：同时属于两个主题的论文从 2023 年的 9 篇增至 2026 年的 437 篇。
• 在「前沿系统」的 4 个主题里，它的 2026 年论文数排第 3；同组增长最快的AI 智能体是 12 倍，推理模型是 3.8 倍。

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

• NeurIPS 2026A Mechanistic Analysis of Looped Reasoning Language Models
• ICML 2026A Formal Comparison Between Chain of Thought and Latent Thought
• ICLR 2026A Simple "Motivation" Can Enhance Reinforcement Finetuning of Large Reasoning Models
• CVPR 2026Asking like Socrates Socrates helps VLMs understand remote sensing images
• NeurIPS 2025A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90% Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1
• ICML 2025A Reasoning-Based Approach to Cryptic Crossword Clue Solving
• ICLR 2025A Formal Framework for Understanding Length Generalization in Transformers
• CVPR 2025Argus: Vision-Centric Reasoning with Grounded Chain-of-Thought
• NeurIPS 2024Are Large Language Models Good Statisticians?
• ICML 2024Algorithm of Thoughts: Enhancing Exploration of Ideas in Large Language Models
• ICLR 2024Amortizing intractable inference in large language models
• CVPR 2024Generate Subgoal Images before Act: Unlocking the Chain-of-Thought Reasoning in Diffusion Model for Robot Manipulation with Multimodal Prompts
• NeurIPS 2023A benchmark of categorical encoders for binary classification
• ICML 2023Can We Scale Transformers to Predict Parameters of Diverse ImageNet Models?
• ICLR 2023Automatic Chain of Thought Prompting in Large Language Models
• CVPR 2023ActMAD: Activation Matching To Align Distributions for Test-Time-Training
• NeurIPS 2026A Solvable Model of Chain-of-Thought in In-Context Learning
• ICML 2026A Narrowing Geometry in Contaminated Reasoning
• ICLR 2026A State-Transition Framework for Efficient LLM Reasoning
• CVPR 2026Beyond Weak Supervision MLLMs-Guided Graded Knowledge Distillation for Unsupervised Camouflaged

全部推理模型论文

共 159 篇，第 2 / 4 页

51. ICML 2025GuardAgent: Safeguard LLM Agents via Knowledge-Enabled Reasoning
52. ICML 2025How Transformers Learn Regular Language Recognition: A Theoretical Study on Training Dynamics and Implicit Bias
53. ICML 2025HyperIV: Real-time Implied Volatility Smoothing
54. ICML 2025HyperTree Planning: Enhancing LLM Reasoning via Hierarchical Thinking
55. ICML 2025Hypo3D: Exploring Hypothetical Reasoning in 3D
56. ICML 2025Imagine While Reasoning in Space: Multimodal Visualization-of-Thought
57. ICML 2025Implicit Language Models are RNNs: Balancing Parallelization and Expressivity
58. ICML 2025Improving Rationality in the Reasoning Process of Language Models through Self-playing Game
59. ICML 2025Interchangeable Token Embeddings for Extendable Vocabulary and Alpha-Equivalence
60. ICML 2025Is Complex Query Answering Really Complex?
61. ICML 2025IT$^3$: Idempotent Test-Time Training
62. ICML 2025KernelBench: Can LLMs Write Efficient GPU Kernels?
63. ICML 2025Large Language Models to Diffusion Finetuning
64. ICML 2025Latent Thought Models with Variational Bayes Inference-Time Computation
65. ICML 2025Learning to (Learn at Test Time): RNNs with Expressive Hidden States
66. ICML 2025Learning to Plan & Reason for Evaluation with Thinking-LLM-as-a-Judge
67. ICML 2025LLMs Can Reason Faster Only If We Let Them
68. ICML 2025LLMs can see and hear without any training
69. ICML 2025LMAct: A Benchmark for In-Context Imitation Learning with Long Multimodal Demonstrations
70. ICML 2025Lower Bounds for Chain-of-Thought Reasoning in Hard-Attention Transformers
71. ICML 2025MA-LoT: Model-Collaboration Lean-based Long Chain-of-Thought Reasoning enhances Formal Theorem Proving
72. ICML 2025Machine Learning meets Algebraic Combinatorics: A Suite of Datasets Capturing Research-level Conjecturing Ability in Pure Mathematics
73. ICML 2025Machines and Mathematical Mutations: Using GNNs to Characterize Quiver Mutation Classes
74. ICML 2025MapEval: A Map-Based Evaluation of Geo-Spatial Reasoning in Foundation Models
75. ICML 2025MARGE: Improving Math Reasoning with Guided Exploration
76. ICML 2025MATH-Perturb: Benchmarking LLMs' Math Reasoning Abilities against Hard Perturbations
77. ICML 2025MathConstruct: Challenging LLM Reasoning with Constructive Proofs
78. ICML 2025Metastable Dynamics of Chain-of-Thought Reasoning: Provable Benefits of Search, RL and Distillation
79. ICML 2025Mind Your Step (by Step): Chain-of-Thought can Reduce Performance on Tasks where Thinking Makes Humans Worse
80. ICML 2025Mirror, Mirror of the Flow: How Does Regularization Shape Implicit Bias?
81. ICML 2025MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and Efficiency
82. ICML 2025Neural Genetic Search in Discrete Spaces
83. ICML 2025Noise Conditional Variational Score Distillation
84. ICML 2025On the Power of Context-Enhanced Learning in LLMs
85. ICML 2025One Example Shown, Many Concepts Known! Counterexample-Driven Conceptual Reasoning in Mathematical LLMs
86. ICML 2025Optimizing Test-Time Compute via Meta Reinforcement Finetuning
87. ICML 2025PDE-Controller: LLMs for Autoformalization and Reasoning of PDEs
88. ICML 2025PENCIL: Long Thoughts with Short Memory
89. ICML 2025PokéChamp: an Expert-level Minimax Language Agent
90. ICML 2025Policy Guided Tree Search for Enhanced LLM Reasoning
91. ICML 2025Position: A Theory of Deep Learning Must Include Compositional Sparsity
92. ICML 2025Position: Formal Mathematical Reasoning—A New Frontier in AI
93. ICML 2025Position: It Is Time We Test Neural Computation In Vitro
94. ICML 2025Position: Language model developers should report train-test overlap
95. ICML 2025Position: LLMs Need a Bayesian Meta-Reasoning Framework for More Robust and Generalizable Reasoning
96. ICML 2025Position: Trustworthy AI Agents Require the Integration of Large Language Models and Formal Methods
97. ICML 2025Position: We Need An Algorithmic Understanding of Generative AI
98. ICML 2025Premise-Augmented Reasoning Chains Improve Error Identification in Math reasoning with LLMs
99. ICML 2025Primal-Dual Neural Algorithmic Reasoning
100. ICML 2025Product of Experts with LLMs: Boosting Performance on ARC Is a Matter of Perspective

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。