mlsanityData to 7 Oct 2026

Recursive Language Models

arXiv 2512.24601 ↗First posted 31 Dec 2025cs.AI cs.CLcited by 94

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

We study allowing large language models (LLMs) to process arbitrarily long prompts through the lens of inference-time scaling. We propose Recursive Language Models (RLMs), a general inference paradigm that treats long prompts as part of an external environment and allows the LLM to programmatically examine, decompose, and recursively call itself over snippets of the prompt. We find that RLMs can successfully process inputs up to two orders of magnitude beyond model context windows and, even for shorter prompts, dramatically outperform the quality of vanilla frontier LLMs and common long-context and coding scaffolds (e.g., on GPT-5 by a median across the evaluated benchmarks of 26% against compaction, 130% against CodeAct with sub-calls, and 13% against Claude Code) across four diverse long-context tasks while having comparable cost. At a small scale, we post-train the first model around the RLM. Our model, RLM-Qwen3-8B, outperforms the underlying Qwen3-8B model by 28.3% on average and even approaches the quality of vanilla GPT-5 on three long-context tasks. Code is available at https://github.com/alexzhang13/rlm.

Key facts. Cited by 94 papers in the map (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 28 of its 60 references that are in the map. First posted on arXiv 31 Dec 2025.

Builds on

28 of 60 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025Self-Steering Language Models9 Apr
  2. 2024THREAD: Thinking Deeper with Recursive Spawning27 May

Other references · 26

  1. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,193
  2. 2018YOLOv3: An Incremental Improvementcited by 2,652
  3. 2021Efficiently Modeling Long Sequences with Structured State Spacescited by 2,424
  4. 2024RULER: What's the Real Context Size of Your Long-Context Language Models?cited by 1,270
  5. 2021Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolationcited by 1,085
  6. 2024Large Language Model based Multi-Agents: A Survey of Progress and Challengescited by 752
  7. 2025Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memorycited by 722
  8. 2023ViperGPT: Visual Inference via Python Execution for Reasoningcited by 591
  9. 2024Executable Code Actions Elicit Better LLM Agentscited by 537
  10. 2024LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitaskscited by 367
  11. 2024Quiet-STaR: Language Models Can Teach Themselves to Think Before Speakingcited by 292
  12. 2021Recursively Summarizing Books with Human Feedbackcited by 287
14 more
  1. 2025MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agentcited by 235
  2. 2024Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attentioncited by 194
  3. 2025BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agentcited by 183
  4. 2023BooookScore: A systematic exploration of book-length summarization in the era of LLMscited by 137
  5. 2025Scaling Long-Horizon LLM Agent via Context-Foldingcited by 136
  6. 2025ReSum: Unlocking Long-Horizon Search Intelligence via Context Summarizationcited by 132
  7. 2023Walking Down the Memory Maze: Beyond Context Limit through Interactive Readingcited by 129
  8. 2025G-Memory: Tracing Hierarchical Memory for Multi-Agent Systemscited by 112
  9. 2025AgentFold: Long-Horizon Web Agents with Proactive Context Managementcited by 95
  10. 2024Michelangelo: Long Context Evaluations Beyond Haystacks via Latent Structure Queriescited by 85
  11. 2021Baleen: Robust Multi-Hop Reasoning at Scale via Condensed Retrievalcited by 63
  12. 2025Oolong: Evaluating Long Context Reasoning and Aggregation Capabilitiescited by 33
  13. 2024Is It Really Long Context if All You Need Is Retrieval? Towards Genuinely Difficult Long Context NLPcited by 27
  14. 2024ReDel: A Toolkit for LLM-Powered Recursive Multi-Agent Systemscited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 94.

94 papers.

Sort:

By area:

  1. 2026LoopFormer: Elastic-Depth Looped Transformers for Latent Reasoning via Shortcut Modulationcited by 48
  2. 2026Meta Context Engineering via Agentic Skill Evolutioncited by 42
  3. 2026τ-Knowledge: Evaluating Conversational Agents over Unstructured Knowledgecited by 35
  4. 2026Continual Harness: Online Adaptation for Self-Improving Foundation Agentscited by 30
  5. 2026HealthBench Professional: Evaluating Large Language Models on Real Clinician Chatscited by 22
  6. 2026Agentic Aggregation for Parallel Scaling of Long-Horizon Agentic Taskscited by 18
  7. 2026RHO: Your Coding Agent is Secretly a Roboticistcited by 18
  8. 2026Coding Agents are Effective Long-Context Processorscited by 11
  9. 2026SideQuest: Model-Driven KV Cache Management for Long-Horizon Agentic Reasoningcited by 10
  10. 2026Recursive Multi-Agent Systemscited by 10
  11. 2026JIT-Agent: Scaling Harness Intelligence via Just-in-Time Harness Evolutioncited by 10
  12. 2026Are Your Reasoning Models Reasoning or Guessing? A Mechanistic Analysis of Hierarchical Reasoning Modelscited by 9
38 more
  1. 2026Recursive Models for Long-Horizon Reasoningcited by 9
  2. 2026PRO-LONG: Programmatic Memory Enables Long-Horizon Reasoningcited by 9
  3. 2026Understanding the Challenges in Iterative Generative Optimization with LLMscited by 8
  4. 2026Workspace Optimization: How to Train Your Agentcited by 7
  5. 2026FutureSim: Replaying World Events to Evaluate Adaptive Agentscited by 7
  6. 2026Decentralized Multi-Agent Systems with Shared Contextcited by 7
  7. 2026Better Harnesses, Smaller Models: Building 90% Cheaper Agents via Automated Harness Adaptationcited by 7
  8. 2026PM-Bench: Evaluating Prospective Memory in LLM Agentscited by 6
  9. 2026Long Context, Less Focus: A Scaling Gap in LLMs Revealed through Privacy and Personalizationcited by 5
  10. 2026Terminal Agents Suffice for Enterprise Automationcited by 5
  11. 2026LongCoT: Benchmarking Long-Horizon Chain-of-Thought Reasoningcited by 5
  12. 2026On Training Large Language Models for Long-Horizon Tasks: An Empirical Study of Horizon Lengthcited by 5
  13. 2026PEEK: Context Map as an Orientation Cache for Long-Context LLM Agentscited by 4
  14. 2026Masking Stale Observations Helps Search Agents -- Until It Doesn't: A Regime Map and Its Mechanismcited by 4
  15. 2026Covert Influence Between Language Modelscited by 4
  16. 2026When Does Continual Learning Require Learningcited by 4
  17. 2026Twin: Playing an Unknown Game with a Test-Time Digital Twincited by 4
  18. 2026Scaling Discovery through Test-Time Communicationcited by 4
  19. 2026ROMA: Recursive Open Meta-Agent Framework for Long-Horizon Multi-Agent Systemscited by 3
  20. 2026Supervising Ralph Wiggum: Exploring a Metacognitive Co-Regulation Agentic AI Loop for Engineering Designcited by 3
  21. 2026Reasoning Shift: How Context Silently Shortens LLM Reasoningcited by 3
  22. 2026Robust Reasoning Benchmarkcited by 3
  23. 2026LCM: Lossless Context Managementcited by 3
  24. 2026SEER: Long-Context Reasoning via Selective Visual-Text Compressioncited by 3
  25. 2026Adaptive Protein Tokenizationcited by 2
  26. 2026DySCO: Dynamic Attention-Scaling Decoding for Long-Context Language Modelscited by 2
  27. 2026Think, But Don't Overthink: Reproducing Recursive Language Modelscited by 2
  28. 2026Deep Reasoning in General Purpose Agents via Structured Meta-Cognitioncited by 2
  29. 2026Multi-Agent Computer Usecited by 2
  30. 2026Declarative Skills for AI Agents in Knowledge-Grounded Tool-Use Workflowscited by 2
  31. 2026Context Compaction Theorycited by 2
  32. 2026Context as an Environment: Programmatic Context Management for Long-Horizon Agentscited by 2
  33. 2026Prefix Sliding for efficient test-time scalingcited by 2
  34. 2026Multi-Agent Systems Should be Treated as Principal-Agent Problemscited by 1
  35. 2026TS-Haystack: A Multi-Task Retrieval Benchmark for Long-Context Time-Series Reasoningcited by 1
  36. 2026Recursive Language Models Meet Uncertainty: The Surprising Effectiveness of Self-Reflective Program Search for Long Contextcited by 1
  37. 2026VideoAtlas: Navigating Long-Form Video in Logarithmic Computecited by 1
  38. 2026Context Cartography: Toward Structured Governance of Contextual Space in Large Language Model Systemscited by 1

50 most cited of 94 All 94 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2023Batch Prompting: Efficient Inference with Large Language Model APIs
  2. 2025ReCAP: Recursive Context-Aware Reasoning and Planning for Large Language Model Agents
  3. 2025Dynamic Compressing Prompts for Efficient Inference of Large Language Models
  4. 2026Prompt Minimization: Reducing Input Redundancy Without Sacrificing Output Fidelity
  5. 2023Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline
  6. 2023Re-Reading Improves Reasoning in Large Language Models
  7. 2025PromptCoT 2.0: Scaling Prompt Synthesis for Large Language Model Reasoning
  8. 2025Learning a Single Token to Replace Long System Prompts in LLMs