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 582 篇，第 2 / 12 页

51. ICML 2026Beyond Explicit Edges: Robust Reasoning over Noisy and Sparse Knowledge Graphs
52. ICML 2026Beyond Majority Voting: Self-Reflective Test-Time Reinforcement Learning for LLM Reasoning
53. ICML 2026Beyond Normalization: Rethinking the Partition Function as a Difficulty Scheduler for RLVR
54. ICML 2026Beyond Scalars: Evaluating and Understanding LLM Reasoning via Geometric Progress and Stability
55. ICML 2026Beyond Static Pipelines: Learning Dynamic Workflows for Text-to-SQL
56. ICML 2026Beyond Test-Time Memory: State-Space Optimal Control for LLM Reasoning
57. ICML 2026Biases in the Blind Spot: Detecting What LLMs Fail to Mention
58. ICML 2026BizFinBench.v2: Towards Reliable LLMs in Finance via Real-User Data and Offline/Online Bilingual Evaluation
59. ICML 2026Blending Supervised and Reinforcement Fine-Tuning with Prefix Sampling
60. ICML 2026Bridging the Grounding Gap in VideoQA via Typed Memory for Language-based Belief-State Reasoning
61. ICML 2026Cache Coherent Resampling for Efficient Test Time Scaling in LLM Reasoning via Adaptive Sequential Monte Carlo
62. ICML 2026CALM Before the STORM: Unlocking Native Reasoning for Optimization Modeling
63. ICML 2026Can LLMs Reason Like Automated Theorem Provers for Rust Verification? VCoT-Bench: Evaluating via Verification Chain of Thought
64. ICML 2026Can LLMs Reason Structurally? Benchmarking via the lens of Data Structures
65. ICML 2026Causal Dependency-Aware Unsupervised Routing for Large Reasoning Models
66. ICML 2026CausalArmor: Efficient Indirect Prompt Injection Guardrails via Causal Attribution
67. ICML 2026CausalProfiler: Generating Synthetic Benchmarks for Rigorous and Transparent Evaluation of Causal Machine Learning
68. ICML 2026CausalXRL: Explainable Reinforcement Learning through Causal Graph Reasoning
69. ICML 2026CauSciBench: Can LLMs Automate Causal Inference in Real-World Scientific Research?
70. ICML 2026Certain Head, Uncertain Tail: Expert-Sample for Test-Time Scaling in Fine-Grained MoE
71. ICML 2026CFPO: Counterfactual Policy Optimization for Multimodal Reasoning
72. ICML 2026Chain-of-Glimpse: Search-Guided Progressive Object-Grounded Reasoning for Video Understanding
73. ICML 2026Chain-of-Goals Hierarchical Policy for Long-Horizon Offline Goal-Conditioned RL
74. ICML 2026Chain-of-Thought Gradient Descent
75. ICML 2026Chain-of-Thought Reasoning In The Wild Is Not Always Faithful
76. ICML 2026Characterizing, Evaluating, and Optimizing Complex Reasoning
77. ICML 2026Chasing Moving Targets with Online Self-Play Reinforcement Learning for Safer Language Models
78. ICML 2026ClinTutor-R1: Advancing Scalable and Robust One-to-Many Alignment in Clinical Socratic Education
79. ICML 2026Clipping Bottleneck: Stabilizing RLVR via Stochastic Recovery of Near-Boundary Signals
80. ICML 2026Closing the Loop: Universal Repository Representation with RPG-Encoder
81. ICML 2026Clustering as Reasoning: A $k$-Means Interpretation of Chain-of-Thought Graph Learning
82. ICML 2026CoCoReviewBench: A Completeness- and Correctness-Oriented Benchmark for AI Reviewers
83. ICML 2026CodeChemist: Test-Time Scaling for Low-Resource Code Generation via Functional Knowledge Transfer
84. ICML 2026CodeClash: Benchmarking Goal-Oriented Software Engineering
85. ICML 2026Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasoner
86. ICML 2026CofactGVR: Counterfactual Intervention for Grounded Visual Reasoning
87. ICML 2026COFT: Counterfactual–Conformal Decoding for Fair Chain‑of‑Thought Reasoning in Large Language Models
88. ICML 2026CoME: Empowering Channel-of-Mobile-Experts with Informative Hybrid-Capabilities Reasoning
89. ICML 2026Compile to Compress: Boosting Formal Theorem Provers by Compiler Outputs
90. ICML 2026Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision
91. ICML 2026Conformal Thinking: Risk Control for Reasoning on a Compute Budget
92. ICML 2026CONGA:Confidence-and-Gradient-Aware Learning Rate Schedule for Test Time Adaptation
93. ICML 2026ConPress: Learning Efficient Reasoning from Multi-Question Contextual Pressure
94. ICML 2026Context-Aware Reasoner: Enhancing Contextual Reasoning in Multimodal Large Language Models
95. ICML 2026Contrastive Reasoning Alignment: Reinforcement Learning from Hidden Representations
96. ICML 2026ConvexBench: Can LLMs Recognize Convex Functions?
97. ICML 2026CoRe: Collaborative Reasoning via Cross Teaching
98. ICML 2026Correcting in Hindsight: Editing Past Key-Value States for Robust LLM Reasoning
99. ICML 2026CoT is Not the Chain of Truth: An Empirical Internal Analysis of Reasoning LLMs for Fake News Generation
100. ICML 2026Counterfactual Residual Data Augmentation for Regression

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。