mlsanityData to 14 Oct 2026

ReasonIR: Training Retrievers for Reasoning Tasks

arXiv 2504.20595 ↗First posted 29 Apr 2025cs.AI cs.CL cs.IRcited by 77

On the map: LLM reasoning → Knowledge- and tool-augmented reasoning

Abstract (arXiv)

We present ReasonIR-8B, the first retriever specifically trained for general reasoning tasks. Existing retrievers have shown limited gains on reasoning tasks, in part because existing training datasets focus on short factual queries tied to documents that straightforwardly answer them. We develop a synthetic data generation pipeline that, for each document, our pipeline creates a challenging and relevant query, along with a plausibly related but ultimately unhelpful hard negative. By training on a mixture of our synthetic data and existing public data, ReasonIR-8B achieves a new state-of-the-art of 29.9 nDCG@10 without reranker and 36.9 nDCG@10 with reranker on BRIGHT, a widely-used reasoning-intensive information retrieval (IR) benchmark. When applied to RAG tasks, ReasonIR-8B improves MMLU and GPQA performance by 6.4% and 22.6% respectively, relative to the closed-book baseline, outperforming other retrievers and search engines. In addition, ReasonIR-8B uses test-time compute more effectively: on BRIGHT, its performance consistently increases with longer and more information-rich rewritten queries; it continues to outperform other retrievers when combined with an LLM reranker. Our training recipe is general and can be easily extended to future LLMs; to this end, we open-source our code, data, and model.

Key facts. Cited by 77 papers in the map (data through 14 Oct 2026). Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning. Builds on 56 of its 65 references that are in the map. First posted on arXiv 29 Apr 2025.

Builds on

56 of 65 references.

Closest ancestors · 9

Same direction, newest first

  1. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning12 Mar
  2. 2025Rank-R1: Enhancing Reasoning in LLM-based Document Rerankers via Reinforcement Learning8 Mar
  3. 2025R1-Searcher: Incentivizing the Search Capability in LLMs via Reinforcement Learning7 Mar
  4. 2025Rank1: Test-Time Compute for Reranking in Information Retrieval25 Feb
  5. 2025Chain-of-Retrieval Augmented Generation24 Jan
  6. 2025Search-o1: Agentic Search-Enhanced Large Reasoning Models9 Jan
  7. 2024RAR-b: Reasoning as Retrieval Benchmark9 Apr
  8. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions20 Dec
  9. 2022Can Retriever-Augmented Language Models Reason? The Blame Game Between the Retriever and the Language Model18 Dec

Other references · 47

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  3. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  4. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  5. 2020Scaling Laws for Neural Language Modelscited by 6,586
  6. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,148
  7. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  8. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,955
  9. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  10. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,724
  11. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameterscited by 1,986
  12. 2016MS MARCO: A Human Generated MAchine Reading COmprehension Datasetcited by 1,605
35 more
  1. 2022Least-to-Most Prompting Enables Complex Reasoning in Large Language Modelscited by 1,407
  2. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,375
  3. 2021Improving language models by retrieving from trillions of tokenscited by 1,293
  4. 2021Unsupervised Dense Information Retrieval with Contrastive Learningcited by 1,216
  5. 2023Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizescited by 737
  6. 2023Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Modelscited by 698
  7. 2020Contrastive Learning with Hard Negative Samplescited by 628
  8. 2023C-Pack: Packed Resources For General Chinese Embeddingscited by 578
  9. 2020Hard Negative Mixing for Contrastive Learningcited by 402
  10. 2023Is ChatGPT Good at Search? Investigating Large Language Models as Re-Ranking Agentscited by 360
  11. 2023Improving Text Embeddings with Large Language Modelscited by 303
  12. 2023Fine-Tuning LLaMA for Multi-Stage Text Retrievalcited by 272
  13. 2024Nomic Embed: Training a Reproducible Long Context Text Embeddercited by 265
  14. 2022Promptagator: Few-shot Dense Retrieval From 8 Examplescited by 235
  15. 2024Generative Representational Instruction Tuningcited by 233
  16. 2020Sparse, Dense, and Attentional Representations for Text Retrievalcited by 214
  17. 2023Query Rewriting for Retrieval-Augmented Large Language Modelscited by 213
  18. 2023Survey on Factuality in Large Language Models: Knowledge, Retrieval and Domain-Specificitycited by 206
  19. 2024BRIGHT: A Realistic and Challenging Benchmark for Reasoning-Intensive Retrievalcited by 158
  20. 2021Scaling Deep Contrastive Learning Batch Size under Memory Limited Setupcited by 127
  21. 2025MMTEB: Massive Multilingual Text Embedding Benchmarkcited by 125
  22. 2021Mr. TyDi: A Multi-lingual Benchmark for Dense Retrievalcited by 119
  23. 2023Retrieval meets Long Context Large Language Modelscited by 115
  24. 2017DuReader: a Chinese Machine Reading Comprehension Dataset from Real-world Applicationscited by 110
  25. 2024LongRAG: Enhancing Retrieval-Augmented Generation with Long-context LLMscited by 108
  26. 2024OpenScholar: Synthesizing Scientific Literature with Retrieval-augmented LMscited by 72
  27. 2024Inference Scaling for Long-Context Retrieval Augmented Generationcited by 65
  28. 2023T2Ranking: A large-scale Chinese Benchmark for Passage Rankingcited by 60
  29. 2024Scaling Retrieval-Based Language Models with a Trillion-Token Datastorecited by 56
  30. 2024Promptriever: Instruction-Trained Retrievers Can Be Prompted Like Language Modelscited by 40
  31. 2024AI-Assisted Generation of Difficult Math Questionscited by 37
  32. 2024CulturalTeaming: AI-Assisted Interactive Red-Teaming for Challenging LLMs' (Lack of) Multicultural Knowledgecited by 28
  33. 2023Instruction Distillation Makes Large Language Models Efficient Zero-shot Rankerscited by 19
  34. 2024BIRCO: A Benchmark of Information Retrieval Tasks with Complex Objectivescited by 15
  35. 2023It's All Relative! -- A Synthetic Query Generation Approach for Improving Zero-Shot Relevance Predictioncited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 38; 2026: 39.

77 papers.

Sort:

By area:

  1. 2025ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memorycited by 284
  2. 2025BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agentcited by 200
  3. 2025Seq vs Seq: An Open Suite of Paired Encoders and Decoderscited by 38
  4. 2025DIVER: A Multi-Stage Approach for Reasoning-intensive Information Retrievalcited by 37
  5. 2025Rank-K: Test-Time Reasoning for Listwise Rerankingcited by 29
  6. 2025RaDeR: Reasoning-aware Dense Retrieval Modelscited by 27
  7. 2025A Comprehensive Survey on Reinforcement Learning-based Agentic Search: Foundations, Roles, Optimizations, Evaluations, and Applicationscited by 22
  8. 2026AgentIR: Reasoning-Aware Retrieval for Deep Research Agentscited by 13
  9. 2025ThinkQE: Query Expansion via an Evolving Thinking Processcited by 12
  10. 2025MRMR: A Realistic and Expert-Level Multidisciplinary Benchmark for Reasoning-Intensive Multimodal Retrievalcited by 12
  11. 2025TongSearch-QR: Reinforced Query Reasoning for Retrievalcited by 11
  12. 2026Rethinking Agentic Search with Pi-Serini: Is Lexical Retrieval Sufficient?cited by 10
38 more
  1. 2026LeanSearch v2: Global Premise Retrieval for Lean 4 Theorem Provingcited by 10
  2. 2025Large Reasoning Embedding Models: Towards Next-Generation Dense Retrieval Paradigmcited by 9
  3. 2026LaSER: Internalizing Explicit Reasoning into Latent Space for Dense Retrievalcited by 9
  4. 2025Frustratingly Simple Retrieval Improves Challenging, Reasoning-Intensive Benchmarkscited by 8
  5. 2025AI-SearchPlanner: Modular Agentic Search via Pareto-Optimal Multi-Objective Reinforcement Learningcited by 8
  6. 2025EviNote-RAG: Enhancing RAG Models via Answer-Supportive Evidence Notescited by 8
  7. 2025Don't "Overthink" Passage Reranking: Is Reasoning Truly Necessary?cited by 7
  8. 2025Retro*: Optimizing LLMs for Reasoning-Intensive Document Retrievalcited by 7
  9. 2025MR2-Bench: Going Beyond Matching to Reasoning in Multimodal Retrievalcited by 7
  10. 2026Revisiting Text Ranking in Deep Researchcited by 7
  11. 2026Procedural Knowledge at Scale Improves Reasoningcited by 7
  12. 2026FutureSim: Replaying World Events to Evaluate Adaptive Agentscited by 7
  13. 2025ERank: Fusing Supervised Fine-Tuning and Reinforcement Learning for Effective and Efficient Text Rerankingcited by 6
  14. 2025Benchmarking Information Retrieval Models on Complex Retrieval Taskscited by 6
  15. 2026TEMPO: A Realistic Multi-Domain Benchmark for Temporal Reasoning-Intensive Retrievalcited by 6
  16. 2026LLMs Should Express Uncertainty Explicitlycited by 6
  17. 2025TFRank: Think-Free Reasoning Enables Practical Pointwise LLM Rankingcited by 5
  18. 2025Beyond Sequential Reranking: Reranker-Guided Search Improves Reasoning Intensive Retrievalcited by 5
  19. 2025PRInTS: Reward Modeling for Long-Horizon Information Seekingcited by 5
  20. 2025Evaluating Legal Reasoning Traces with Legal Issue Tree Rubricscited by 5
  21. 2026Can Large Language Models Keep Up? Benchmarking Online Adaptation to Continual Knowledge Streamscited by 5
  22. 2025InsertRank: LLMs can reason over BM25 scores to Improve Listwise Rerankingcited by 4
  23. 2025Reason to Retrieve: Enhancing Query Understanding through Decomposition and Interpretationcited by 4
  24. 2025LLM-guided Hierarchical Search for End-to-end Reasoning Intensive Retrievalcited by 4
  25. 2026Are LLM-Based Retrievers Worth Their Cost? An Empirical Study of Efficiency, Robustness, and Reasoning Overheadcited by 4
  26. 2026What Limits Does Quantization Place on Dense Top-k Retrieval? A Theoretical Studycited by 4
  27. 2025RAG in the Wild: On the (In)effectiveness of LLMs with Mixture-of-Knowledge Retrieval Augmentationcited by 3
  28. 2025On The Role of Pretrained Language Models in General-Purpose Text Embeddings: A Surveycited by 3
  29. 2025LimRank: Less is More for Reasoning-Intensive Information Rerankingcited by 3
  30. 2026ARK: A Dual-Axis Multimodal Retrieval Benchmark along Reasoning and Knowledgecited by 3
  31. 2025From Token to Action: State Machine Reasoning to Mitigate Overthinking in Information Retrievalcited by 2
  32. 2025A Vision for Geo-Temporal Deep Research Systems: Towards Comprehensive, Transparent, and Reproducible Geo-Temporal Information Synthesiscited by 2
  33. 2026Rethinking Reasoning-Intensive Retrieval: Evaluating and Advancing Retrievers in Agentic Search Systemscited by 2
  34. 2026Scaling Participation in Modular AI Systemscited by 2
  35. 2026RL-Index: Reinforcement Learning for Retrieval Index Reasoningcited by 2
  36. 2026Antares: Foundation Models for Agentic Vulnerability Localizationcited by 2
  37. 2025Lighting the Way for BRIGHT: Reproducible Baselines with Anserini, Pyserini, and RankLLMcited by 1
  38. 2025LIR: The First Workshop on Late Interaction and Multi Vector Retrieval @ ECIR 2026cited by 1

50 most cited of 77 All 77 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Your Dense Retriever is Secretly an Expeditious Reasoner
  2. 2026Multi-Step Semantic Reasoning in Generative Retrieval
  3. 2026When to Retrieve During Reasoning: Adaptive Retrieval for Large Reasoning Models
  4. 2025Retrieval-in-the-Chain: Bootstrapping Large Language Models for Generative Retrieval
  5. 2025Retrieval-augmented reasoning with lean language models
  6. 2025FREESON: Retriever-Free Retrieval-Augmented Reasoning via Corpus-Traversing MCTS
  7. 2025Hybrid Deep Searcher: Scalable Parallel and Sequential Search Reasoning
  8. 2026Retrieval-Infused Reasoning Sandbox: A Benchmark for Decoupling Retrieval and Reasoning Capabilities