Inference-Time Scaling

Follow this topic Ask about this topic

Inference-time scaling asks how models can use additional computation after training to solve harder problems, improve generation, or act more reliably. Common approaches include sampling and self-consistency, search and backtracking, verifier-guided selection, recurrent computation, and adaptive allocation across tasks or agents; they are tested on settings such as math, code, theorem proving, scientific discovery, and embodied planning. Results often improve with more compute, but gains depend on the task, budget, and quality of selection signals, and can come with substantial latency and resource costs. A central tension is deciding when extra computation helps: correlated errors, weak verifiers, or longer erroneous trajectories can make scaling ineffective or even harmful, so evaluations increasingly compare methods at matched budgets.

Latest papers

115 papers27 papers in the last four weeks, up 170% on the four weeks before. 0.2% of all new papers.

Direct Pass@ $k$ Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning
Oct 22, 2026Mahdi Salmani, Adel Javanmard, Devansh Gupta +1Efficient Language Model Reasoning LLM Pruning
University of Southern California
Where Should a Stuck Agent Spend Its Next Tokens? Causal Attribution of Inference-Time Resources
Oct 20, 2026Saiyue Lyu, Jerrold Huang, Yunbei Zhang +4Inference-Time Compute Allocation Software Engineering Agents
University of British Columbia · University of Toronto · Tulane University +2
Test-Time Scaling of Text-to-Image Flow Matching Models with Noise-Rejuvenated Sequential Monte Carlo
Oct 18, 2026Wenyi Lian, Zheng Zhao, Thomas B. SchönInference-Time Scaling T2I Generation
Department of Information Technology Uppsala University, Sweden · Department of Computer and Information Science Linköping University, Sweden
World of SWECraft: Agent Collaboration at Scale via Simulated Software Worlds
Oct 16, 2026Shannon Zejiang Shen, Yapei Chang, Qiyao Liang +13Multi-Agent Collaboration Software Engineering Agents
Massachusetts Institute of Technology · University of Maryland, College Park · University of Washington +3
TIPS: A Tiny Inference-time Policy for Scaling in Diffusion Language Models
Oct 15, 2026Ya-Shi Zhang, Jian Tang, Pierre-André Noël +1Diffusion Language Model Decoding Inference-Time Adaptation of Language Models
ServiceNow AI Research · Mila - Quebec AI Institute · HEC Montréal
Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally
Oct 15, 2026Zhizhuo Kou, Jian Yang, Pengfei Hao +4Pragmatic Reasoning in Language Models Large Reasoning Models
1The Hong Kong University of Science and Technology · 2The Hong Kong University of Science and Technology (Guangzhou) · Peking University
Decoupled Marginal Sharpening for Training-Free Inference-Time Scaling
Oct 14, 2026Weixin Wang, Wei Deng, Anderson Schneider +3Inference-Time Scaling Efficient Language Model Reasoning
Duke University · Morgan Stanley
Tail Competition Explains Scaling in Best-of-N Sampling for Verifiable Problems
Oct 14, 2026Taro Yano, Yoichi Ishibashi, Masafumi OyamadaBest-of-N Sampling Inference-Time Scaling
NEC Corporation
Input-Conditioned Search for Efficient Test-Time Scaling of Recursive Reasoners
Oct 13, 2026Haochen Shi, Tianwen Fu, Qianggang Ding +7Inference-Time Search Inference-Time Scaling
Mila – Quebec AI Institute · Université de Montréal · University of Southern California +2
Scaling Inference-Time Compute with Teams of Coding Agents
Oct 12, 2026Keenan Samway, Daniel D'souza, Omar Darwiche Domingues +3Inference-Time Scaling Inference-Time Compute Allocation
Cohere Labs · Cohere
Recurrent In-Context Learning for Tabular Foundation Models
Oct 10, 2026Tejas Vaidhya, Gaurav Sahu, Irina RishTabular Foundation Models Tabular ICL
Mila – Quebec AI Institute · Nolano AI · Unified Sciences
Test-Time Compute for Tabular Foundation Models: Mechanisms, Gains, and Limits
Oct 8, 2026Kanghui Ning, Marin Biloš, James T. Wilson +5Tabular Foundation Models Inference-Time Scaling
School of Computing, University of Connecticut, Storrs, USA. · Department of Machine Learning Research, Morgan Stanley, New York, USA.
Scaling to Tens of Thousands of Test-Time Iterations with Loop-Native Attention Residuals
Oct 8, 2026Pengxiang Li, Dilxat Muhtar, Di He +3Recurrent Transformers Attention Residual Connections
The Hong Kong Polytechnic University · Independent Researcher · Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences +6
The Long Road to the Same Answer: Cognitive Bias Under Escalating Reasoning Budgets in Large Language Models
Oct 7, 2026Obada KraishanCognitive Bias in Language Models Large Reasoning Models
College of Media and Communication Texas Tech University Lubbock, TX, USA
Expanding LLM Reasoning
Oct 4, 2026Rian Atri, Evan LuoLLM Inference Efficiency Inference-Time Scaling
Keiji AI · University of California, Berkeley
When Does Longer Reasoning Help? Predicting Mathematical Reasoning Through Discovery and Execution
Oct 4, 2026Adib Hasan, Lay Jain, Thanic Nur SaminInference-Time Scaling LLM Mathematical Reasoning
Independent Researcher · IU Bloomington
GitSwarm: Decentralized Compounding Inference
Oct 4, 2026Vedant Shah, Ankur Samanta, Paras Dahal +11Inference-Time Scaling Multi-Agent Collaboration
Meta Superintelligence Labs · Mila - Québec AI Institute · Université de Montréal +4
Error-Corrected Inference-Time Scaling for Imperfect Diffusion Models
Oct 1, 2026Zuokai Wen, Louis Grenioux, Weinan E +1Markov Chain Monte Carlo Diffusion Model Sampling
School of Artificial Intelligence, Shanghai Jiao Tong University · Center for Computational Mathematics, Flatiron Institute · School of Mathematical Sciences, Peking University
Scaling Laws for Looped Mixture of Experts
Sep 30, 2026Yanbei Chen, Anirudh Goyal, Raghuraman KrishnamoorthiMixture-of-Experts Language Models Language Model Scaling Laws
Meta AI
Hermes: Learning Contextual Reasoning Unlocks Test-Time Scaling
Sep 29, 2026Xinyu Li, Mononito Goswami, Hao Liu +5LLM Inference Inference-Time Scaling
Carnegie Mellon University · AWS AI Labs · Washington University in St. Louis +1
Explore Broadly, Reason Sharply: Push Small Models toward the Frontier via Sampling
Sep 29, 2026Panagiotis Theodoropoulos, Nan Jiang, Xintong Duan +4Efficient Language Model Reasoning Parallel Tempering
Georgia Institute of Technology · Morgan Stanley · University of Texas at El Paso +1
Port-Hamiltonian Latent Deliberation: Mitigating the Deliberation Drift Cliff in Test-Time Compute Scaling
Sep 29, 2026Zeyu JiaHodge Decomposition Inference-Time Scaling
School of Biomedical Engineering and Technology, Tianjin Medical University, Medical School, Tianjin University
Scheduling Recursive Reasoning in Looped Transformers
Sep 29, 2026Boyuan Wang, Chengyao Yu, Jiaxi Ren +3Adaptive Inference Recurrent Transformers
Southern University of Science and Technology · The Chinese University of Hong Kong, Shenzhen · Shenzhen Loop Area Institute
What Makes Recurrence Effective in Looped Language Models?
Sep 29, 2026Xinlin Zhuang, Siyuan Wang, Imran Razzak +1Recursive Language Models Efficient Language Model Inference
The Chinese University of Hong Kong · MBZUAI
From Search to Research: Exploring Search Scaling in Autonomous Quantitative Factor Mining
Sep 28, 2026Kangcheng Deng, Hui Cai, Jiacheng Lu +6Inference-Time Scaling Autonomous Research Agents
StepFun · Shanghai Jiao Tong University · University of California, Los Angeles +1
Long-Horizon Scaling: How Model Capabilities Shape the Returns to Computation
Sep 28, 2026Haoyu Zheng, Zhengyu Chen, Huaisheng Zhu +5Cost-Aware Inference Inference-Time Scaling
Meituan LongCat Team · Zhejiang University · Westlake University +2
Compute Time Scaling with Recursive Models for Combinatorial Optimization
Sep 28, 2026Zhengxi Zhang, Paul SwobodaTraveling Salesperson Problem Inference-Time Scaling
Scaling of Capability and Efficiency at Inference Time in Large Reasoning Models
Sep 22, 2026Moritz Laber, Zohair Shafi, Germans Savcisens +6LLM Inference Efficiency Language Model Scaling Laws
Northeastern University, Boston, Massachusetts, USA · Complexity Science Hub Vienna, Vienna, Austria · Santa Fe Institute, Santa Fe, New Mexico, USA
The Organization of Inference: Information, Resource Constraints, and AI Production
Sep 17, 2026Yukun Zhang, Kemu Xu, Yishen ChenLLM Planning Inference-Time Scaling
The Chinese University of Hong Kong Hong Kong, China · University of Edinburgh Edinburgh, United Kingdom · The Chinese University of Hong Kong, Shenzhen Shenzhen, China
Right Direction, Wrong Step: Geometric Analysis of Finite-Step Failure in Looped Transformers
Sep 15, 2026Zhihao Guo, Zonghan Wu, Haizhou Du +4Inference-Time Scaling Recurrent Transformers
University of Technology Sydney, Australia · East China Normal University, China · Shanghai University of Electric Power, China +2

Oct 22, 2026J/K move · Enter open · S save

Direct Pass@ $k$ Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning

Mahdi Salmani, Adel Javanmard, Devansh Gupta, Meisam Razaviyayn

University of Southern California

Efficient Language Model Reasoning LLM Pruning Inference-Time Scaling

Open paper AskPDF arXiv

Test-time scaling has emerged as a key inference strategy for reasoning models, where a model generates a large number of candidate solutions in the hope of finding at least one correct solution. Running all candidate solutions to completion can be computationally demanding, so recent work prunes the candidate pool after generating only a short prefix, committing further computation only to the surviving candidates. Existing pruning strategies divide broadly into two categories: Quality-based methods assign a score to each partial solution and discard those predicted to ultimately lead to an incorrect final solution, while diversity-based methods select partial solutions that follow different reasoning trajectories and are therefore likely to cover different final solutions. The tension between these two objectives is an instance of exploration versus exploitation, yet little to no prior work on pruning exposes this trade-off directly, let alone optimizes it. In this paper, we introduce a pruning procedure that addresses this trade-off directly by optimizing pass@

k

over the selected solution pool. In particular, it chooses the pool to maximize the probability that at least one of its partial solutions, when continued to completion, yields a correct answer. Because this objective simultaneously rewards high-quality partial solutions and penalizes redundancy among them, quality and diversity are naturally balanced by the criterion itself rather than through separate ad-hoc design choices. On eight open-weight reasoning models, our method outperforms existing quality-based and diversity-based pruning strategies. The gains are particularly pronounced when the pruning procedure keeps a moderate number of partial solutions for continued generation, highlighting the value of directly balancing quality and diversity during pruning.