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 652 篇，第 3 / 14 页

101. ICLR 2026DeepCompress: A Dual Reward Strategy for Dynamically Exploring and Compressing Reasoning Chains
102. ICLR 2026DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning
103. ICLR 2026DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search
104. ICLR 2026Demystifying and Enhancing the Efficiency of Large Language Model Based Search Agents
105. ICLR 2026DePO: Demonstration-guided Policy Optimization for Molecular Optimization
106. ICLR 2026DepthLM: Metric Depth from Vision Language Models
107. ICLR 2026DESIGNER: Design-Logic-Guided Multidisciplinary Data Synthesis for LLM Reasoning
108. ICLR 2026Diagnosing and Remedying Knowledge Deficiencies in LLMs via Label-free Curricular Meaningful Learning
109. ICLR 2026DiffAdapt: Difficulty-Adaptive Reasoning for Token-Efficient LLM Inference
110. ICLR 2026Differential Fine-Tuning Large Language Models Towards Better Diverse Reasoning Abilities
111. ICLR 2026Difficulty–Diversity Collaborative Filtering for Data-Efficient LLM Fine-Tuning
112. ICLR 2026Diffusion Language Models are Provably Optimal Parallel Samplers
113. ICLR 2026Disrupting Hierarchical Reasoning: Adversarial Protection for Geographic Privacy in Multimodal Reasoning Models
114. ICLR 2026Distilled Pretraining: A modern lens of Data, In-Context Learning and Test-Time Scaling
115. ICLR 2026Distilling the Thought, Watermarking the Answer: A Principle Semantic Guided Watermark for Reasoning Large Language Models
116. ICLR 2026Diversity-Enhanced Reasoning for Subjective Questions
117. ICLR 2026Diversity-Incentivized Exploration for Versatile Reasoning
118. ICLR 2026Do 3D Large Language Models Really Understand 3D Spatial Relationships?
119. ICLR 2026Do Vision-Language Models Respect Contextual Integrity in Location Disclosure?
120. ICLR 2026Don’t Pass@$k$: A Bayesian Framework for Large Language Model Evaluation
121. ICLR 2026Doxing via the Lens: Revealing Location-related Privacy Leakage on Multi-modal Large Reasoning Models
122. ICLR 2026DRAGON: Guard LLM Unlearning in Context via Negative Detection and Reasoning
123. ICLR 2026Draw-In-Mind: Rebalancing Designer-Painter Roles in Unified Multimodal Models Benefits Image Editing
124. ICLR 2026DRPO: Efficient Reasoning via Decoupled Reward Policy Optimization
125. ICLR 2026DSSA: Dense-Sparse Switchable Attention for Seamless Short-to-Long Adaptation
126. ICLR 2026Dyna-Mind: Learning to Simulate from Experience for Better AI Agents
127. ICLR 2026DynaGuard: A Dynamic Guardian Model With User-Defined Policies
128. ICLR 2026Dynamic Early Exit in Reasoning Models
129. ICLR 2026Dynamics-Predictive Sampling for Active RL Finetuning of Large Reasoning Models
130. ICLR 2026e3: Learning to Explore Enables Extrapolation of Test-Time Compute for LLMs
131. ICLR 2026Early Signs of Steganographic Capabilities in Frontier LLMs
132. ICLR 2026Easier Painting Than Thinking: Can Text-to-Image Models Set the Stage, but Not Direct the Play?
133. ICLR 2026Echo: Towards Advanced Audio Comprehension via Audio-Interleaved Reasoning
134. ICLR 2026Echoes as Anchors: Probabilistic Costs and Attention Refocusing in LLM Reasoning
135. ICLR 2026EEPO: Exploration-Enhanced Policy Optimization via Sample-Then-Forget
136. ICLR 2026Efficient Ensemble Conditional Independence Test Framework for Causal Discovery
137. ICLR 2026Efficient Multimodal Spatial Reasoning via Dynamic and Asymmetric Routing
138. ICLR 2026Efficient Reasoning with Balanced Thinking
139. ICLR 2026Efficient Test-Time Scaling for Small Vision-Language Models
140. ICLR 2026Efficient Turing Machine Simulation with Transformers
141. ICLR 2026Embodied-R1: Reinforced Embodied Reasoning for General Robotic Manipulation
142. ICLR 2026Emergence of Superposition: Unveiling the Training Dynamics of Chain of Continuous Thought
143. ICLR 2026Emergent Discrete Controller Modules for Symbolic Planning in Transformers
144. ICLR 2026Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning
145. ICLR 2026EmotionThinker: Prosody-Aware Reinforcement Learning for Explainable Speech Emotion Reasoning
146. ICLR 2026Energy-Based Transformers are Scalable Learners and Thinkers
147. ICLR 2026Enhancing Complex Symbolic Logical Rea­soning of Large Language Models via Sparse Multi-Agent Debate
148. ICLR 2026Enhancing Language Model Reasoning with Structured Multi-Level Modeling
149. ICLR 2026Enhancing LLMs for Knowledge Base Question Answering by Chain-of-Decomposition
150. ICLR 2026ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models

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。