mlsanityData to 14 Oct 2026

OpenThoughts: Data Recipes for Reasoning Models

arXiv 2506.04178 ↗First posted 4 Jun 2025cs.LGcited by 304

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

Reasoning models have made rapid progress on many benchmarks involving math, code, and science. Yet, there are still many open questions about the best training recipes for reasoning since state-of-the-art models often rely on proprietary datasets with little to no public information available. To address this, the goal of the OpenThoughts project is to create open-source datasets for training reasoning models. After initial explorations, our OpenThoughts2-1M dataset led to OpenThinker2-32B, the first model trained on public reasoning data to match DeepSeek-R1-Distill-32B on standard reasoning benchmarks such as AIME and LiveCodeBench. We then improve our dataset further by systematically investigating each step of our data generation pipeline with 1,000+ controlled experiments, which led to OpenThoughts3. Scaling the pipeline to 1.2M examples and using QwQ-32B as teacher yields our OpenThoughts3-7B model, which achieves state-of-the-art results: 53% on AIME 2025, 51% on LiveCodeBench 06/24-01/25, and 54% on GPQA Diamond - improvements of 15.3, 17.2, and 20.5 percentage points compared to the DeepSeek-R1-Distill-Qwen-7B. All of our datasets and models are available on https://openthoughts.ai.

Key facts. Cited by 304 papers in the map (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 36 of its 66 references that are in the map. First posted on arXiv 4 Jun 2025.

Builds on

36 of 66 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025Skywork Open Reasoner 1 Technical Report28 May
  2. 2025Phi-4-reasoning Technical Report30 Apr

Other references · 34

  1. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  2. 2024Qwen2.5 Technical Reportcited by 4,942
  3. 2020The Pile: An 800GB Dataset of Diverse Text for Language Modelingcited by 2,411
  4. 2024LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Codecited by 2,004
  5. 2024Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonecited by 1,917
  6. 2022Competition-Level Code Generation with AlphaCodecited by 1,462
  7. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,429
  8. 2025s1: Simple test-time scalingcited by 1,341
  9. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!cited by 1,154
  10. 2024Tulu 3: Pushing Frontiers in Open Language Model Post-Trainingcited by 1,053
  11. 2023Gemini: A Family of Highly Capable Multimodal Modelscited by 872
  12. 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMscited by 570
22 more
  1. 2023XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Modelscited by 505
  2. 2025LIMO: Less is More for Reasoningcited by 498
  3. 2025Humanity's Last Examcited by 491
  4. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Modelscited by 464
  5. 2025DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoningcited by 280
  6. 2024OpenMathInstruct-2: Accelerating AI for Math with Massive Open-Source Instruction Datacited by 200
  7. 2025AIMO-2 Winning Solution: Building State-of-the-Art Mathematical Reasoning Models with OpenMathReasoning datasetcited by 174
  8. 2023TACO: Topics in Algorithmic COde generation datasetcited by 160
  9. 2024Nemotron-CC: Transforming Common Crawl into a Refined Long-Horizon Pretraining Datasetcited by 160
  10. 2025Llama-Nemotron: Efficient Reasoning Modelscited by 158
  11. 2024Nemotron-4 340B Technical Reportcited by 144
  12. 2024Rho-1: Not All Tokens Are What You Needcited by 133
  13. 2025OpenVLThinker: Complex Vision-Language Reasoning via Iterative SFT-RL Cyclescited by 124
  14. 2025CodeElo: Benchmarking Competition-level Code Generation of LLMs with Human-comparable Elo Ratingscited by 115
  15. 2025OpenCodeReasoning: Advancing Data Distillation for Competitive Codingcited by 113
  16. 2025NaturalReasoning: Reasoning in the Wild with 2.8M Challenging Questionscited by 84
  17. 2025Organize the Web: Constructing Domains Enhances Pre-Training Data Curationcited by 76
  18. 2024Alice in Wonderland: Simple Tasks Showing Complete Reasoning Breakdown in State-Of-the-Art Large Language Modelscited by 68
  19. 2025STAR-1: Safer Alignment of Reasoning LLMs with 1K Datacited by 68
  20. 2024Autonomous Data Selection with Zero-shot Generative Classifiers for Mathematical Textscited by 27
  21. 2025Predictive Data Selection: The Data That Predicts Is the Data That Teachescited by 21
  22. 2025Datasets, Documents, and Repetitions: The Practicalities of Unequal Data Qualitycited by 8

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 102; 2026: 202.

304 papers.

Sort:

By area:

  1. 2026Reinforcement Learning via Self-Distillationcited by 332
  2. 2026Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipecited by 299
  3. 2025MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipecited by 186
  4. 2026Learning beyond Teacher: Generalized On-Policy Distillation with Reward Extrapolationcited by 184
  5. 2026Entropy-Aware On-Policy Distillation of Language Modelscited by 175
  6. 2025Scaling Latent Reasoning via Looped Language Modelscited by 165
  7. 2026Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixescited by 145
  8. 2025Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoningcited by 106
  9. 2025On-Policy RL Meets Off-Policy Experts: Harmonizing Supervised Fine-Tuning and Reinforcement Learning via Dynamic Weightingcited by 91
  10. 2025MegaScience: Pushing the Frontiers of Post-Training Datasets for Science Reasoningcited by 59
  11. 2025Black-Box On-Policy Distillation of Large Language Modelscited by 59
  12. 2025RLVE: Scaling Up Reinforcement Learning for Language Models with Adaptive Verifiable Environmentscited by 50
38 more
  1. 2025Klear-Reasoner: Advancing Reasoning Capability via Gradient-Preserving Clipping Policy Optimizationcited by 39
  2. 2025MMR1: Enhancing Multimodal Reasoning with Variance-Aware Sampling and Open Resourcescited by 39
  3. 2025OpenMMReasoner: Pushing the Frontiers for Multimodal Reasoning with an Open and General Recipecited by 37
  4. 2026OpenResearcher: A Fully Open Pipeline for Long-Horizon Deep Research Trajectory Synthesiscited by 37
  5. 2026Asymmetric On-Policy Distillation: Bridging Exploitation and Imitation at the Token Levelcited by 35
  6. 2025RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMscited by 34
  7. 2026Lightning OPD: Efficient Post-Training for Large Reasoning Models with Offline On-Policy Distillationcited by 33
  8. 2025NaturalThoughts: Selecting and Distilling Reasoning Traces for General Reasoning Taskscited by 32
  9. 2025Reverse-Engineered Reasoning for Open-Ended Generationcited by 32
  10. 2026MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methodscited by 30
  11. 2026When Are Teacher Tokens Reliable? Position-Weighted On-Policy Self-Distillation for Reasoningcited by 30
  12. 2026Trust Region On-Policy Distillationcited by 30
  13. 2025Quagmires in SFT-RL Post-Training: When High SFT Scores Mislead and What to Use Insteadcited by 28
  14. 2025MedVLThinker: Simple Baselines for Multimodal Medical Reasoningcited by 27
  15. 2025Front-Loading Reasoning: The Synergy between Pretraining and Post-Training Datacited by 27
  16. 2025Cognitive Foundations for Reasoning and Their Manifestation in LLMscited by 26
  17. 2025Wait, Wait, Wait... Why Do Reasoning Models Loop?cited by 26
  18. 2026Reasoning Models Struggle to Control their Chains of Thoughtcited by 25
  19. 2025RL Grokking Recipe: How Does RL Unlock and Transfer New Algorithms in LLMs?cited by 23
  20. 2026Respecting Self-Uncertainty in On-Policy Self-Distillation for Efficient LLM Reasoningcited by 23
  21. 2026Olmo Hybrid: From Theory to Practice and Backcited by 22
  22. 2025KnowRL: Exploring Knowledgeable Reinforcement Learning for Factualitycited by 20
  23. 2026Decoupling KL and Trajectories: A Unified Perspective for SFT, DAgger, Offline RL, and OPD in LLM Distillationcited by 20
  24. 2025Plan Then Action:High-Level Planning Guidance Reinforcement Learning for LLM Reasoningcited by 19
  25. 2025Trace Length is a Simple Uncertainty Signal in Reasoning Modelscited by 19
  26. 2026WebGym: Scaling Training Environments for Visual Web Agents with Realistic Taskscited by 19
  27. 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
  28. 2025Language Models Can Learn from Verbal Feedback Without Scalar Rewardscited by 18
  29. 2026DMax: Aggressive Parallel Decoding for dLLMscited by 18
  30. 2025Can We Predict Alignment Before Models Finish Thinking? Towards Monitoring Misaligned Reasoning Modelscited by 17
  31. 2025ARES: Multimodal Adaptive Reasoning via Difficulty-Aware Token-Level Entropy Shapingcited by 17
  32. 2025AgentMath: Empowering Mathematical Reasoning for Large Language Models via Tool-Augmented Agentcited by 17
  33. 2026PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoningcited by 17
  34. 2026Distribution-Aligned Sequence Distillation for Superior Long-CoT Reasoningcited by 17
  35. 2026MAD-OPD: Breaking the Ceiling in On-Policy Distillation via Multi-Agent Debatecited by 17
  36. 2026On-Policy Self-Distillation without Any Supervisioncited by 17
  37. 2025Proximal Supervised Fine-Tuningcited by 16
  38. 2025QeRL: Beyond Efficiency -- Quantization-enhanced Reinforcement Learning for LLMscited by 16

50 most cited of 304 All 304 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Reasoning with OmniThought: A Large CoT Dataset with Verbosity and Cognitive Difficulty Annotations
  2. 2025One Missing Piece for Open-Source Reasoning Models: A Dataset to Mitigate Cold-Starting Short CoT LLMs in RL
  3. 2025Not All Correct Answers Are Equal: Why Your Distillation Source Matters
  4. 2025A Short Survey on Small Reasoning Models: Training, Inference, Applications and Research Directions
  5. 2024A Comparative Study on Reasoning Patterns of OpenAI's o1 Model
  6. 2025Learning to Reason: Training LLMs with GPT-OSS or DeepSeek R1 Reasoning Traces
  7. 2025ReasoningWeekly: A General Knowledge and Verbal Reasoning Challenge for Large Language Models
  8. 2024OpenR: An Open Source Framework for Advanced Reasoning with Large Language Models