mlsanityData to 7 Oct 2026

Autodata: An agentic data scientist to create high quality synthetic data

arXiv 2606.25996 ↗First posted 24 Jun 2026cs.AI cs.CL cs.LGcited by 12

On the map: LLM agents and tool use → Web, search, and research agents

Abstract (arXiv)

We introduce Autodata, a general method that enables AI agents to act as data scientists who build high quality training and evaluation data. We show how to train (meta-optimize) such a data scientist agent, so that it learns to create even stronger data. We describe the overall formulation, and a specific practical implementation, Agentic Self-Instruct. We conduct experiments on computer science research tasks, legal reasoning tasks and reasoning with mathematical objects, where we obtain improved results compared to classical synthetic dataset creation methods. Further, meta-optimizing the data scientist agent itself delivers an even larger performance uplift. Agentic data creation provides a way to convert increased inference compute into higher quality model training. Overall, we believe this direction has the potential to change the way we build AI data.

Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Placed in Web, search, and research agents, one of 16 directions in LLM agents and tool use. Builds on 35 of its 38 references that are in the map. First posted on arXiv 24 Jun 2026.

Builds on

35 of 38 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025AutoData: A Multi-Agent System for Open Web Data Collection21 May
  2. 2024Data Interpreter: An LLM Agent For Data Science28 Feb
  3. 2024DS-Agent: Automated Data Science by Empowering Large Language Models with Case-Based Reasoning27 Feb

Other references · 32

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  2. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,317
  3. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
  4. 2023WizardLM: Empowering large pre-trained language models to follow complex instructionscited by 1,116
  5. 2024The AI Scientist: Towards Fully Automated Open-Ended Scientific Discoverycited by 1,002
  6. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 895
  7. 2023Large Language Models as Optimizerscited by 830
  8. 2023Enhancing Chat Language Models by Scaling High-quality Instructional Conversationscited by 825
  9. 2023MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Modelscited by 724
  10. 2024Self-Rewarding Language Modelscited by 654
  11. 2023Textbooks Are All You Need II: phi-1.5 technical reportcited by 559
  12. 2023Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolutioncited by 521
20 more
  1. 2023MAmmoTH: Building Math Generalist Models through Hybrid Instruction Tuningcited by 483
  2. 2025GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningcited by 449
  3. 2023Orca: Progressive Learning from Complex Explanation Traces of GPT-4cited by 369
  4. 2024Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothingcited by 341
  5. 2025Absolute Zero: Reinforced Self-play Reasoning with Zero Datacited by 323
  6. 2026Meta-Harness: End-to-End Optimization of Model Harnessescited by 248
  7. 2024OpenMathInstruct-2: Accelerating AI for Math with Massive Open-Source Instruction Datacited by 194
  8. 2022Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Datasetcited by 113
  9. 2024AgentInstruct: Toward Generative Teaching with Agentic Flowscited by 88
  10. 2025NaturalReasoning: Reasoning in the Wild with 2.8M Challenging Questionscited by 81
  11. 2025SPICE: Self-Play In Corpus Environments Improves Reasoningcited by 78
  12. 2025Self-Challenging Language Model Agentscited by 56
  13. 2025The Majority is not always right: RL training for solution aggregationcited by 46
  14. 2024AI-Assisted Generation of Difficult Math Questionscited by 37
  15. 2025CoT-Self-Instruct: Building high-quality synthetic prompts for reasoning and non-reasoning taskscited by 30
  16. 2025PRBench: Large-Scale Expert Rubrics for Evaluating High-Stakes Professional Reasoningcited by 29
  17. 2024Source2Synth: Synthetic Data Generation and Curation Grounded in Real Data Sourcescited by 27
  18. 2024DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedbackcited by 23
  19. 2025AI & Human Co-Improvement for Safer Co-Superintelligencecited by 8
  20. 2026Reasoning over mathematical objects: on-policy reward modeling and test time aggregationcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 12.

12 papers.

  1. 2026Recursive Synthesis for Long-Horizon Terminal Taskscited by 8
  2. 2026SPADE: Self-Play in Adaptive Synthetic Executable Environmentscited by 6
  3. 2026Recursive self-improvement of AI research agentscited by 3
  4. 2026Execution-First Synthetic Tool-Use Trace Generation for LLM Agentscited by 1
  5. 2026Scaling Automatic Research Agents via World Modelscited by 1
  6. 2026What is Missing from AI Post-Training AI: An Empirical Analysiscited by 1
  7. 2026A Vocabulary for Multi-Agent Automated Research Systemscited by 0
  8. 2026RAVEN-Eval: Rubric-Guided Automatic Evaluation for AI Video Generation Models Based on LMM Preference Judgementcited by 0
  9. 2026DataFoundry: Evolving Data Preparators via Recursive Self-Improvementcited by 0
  10. 2026Alignment Forecasting: Predicting Misalignment From Training Datacited by 0
  11. 2026G"odel Forest: Balancing Search Depth and Breadth for Data-Centric Recursive Self-Improvementcited by 0
  12. 2026Invent a Dataset: Measuring dataset generation abilities with zero seedcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Reasoning-Driven Synthetic Data Generation and Evaluation
  2. 2026AutoData: Agentic Search for Pre-training Data Selection
  3. 2025Can Agentic AI Match the Performance of Human Data Scientists?
  4. 2026Can Generalist Agents Automate Data Curation?
  5. 2026AutoDataBench: Can Agents Write the Data That Feeds the Self-Improvement Loop?
  6. 2026AgenticDataBench: A Comprehensive Benchmark for Data Agents
  7. 2026Data Agents: Agentic Data Systems
  8. 2025AutoMind: Adaptive Knowledgeable Agent for Automated Data Science