mlsanityData to 7 Oct 2026

Web, search, and research agents

1,671 papers14 this weekLow confidence in this grouping: the papers may belong to more than one topic, or the group may changeLow confidence in this grouping: the papers may belong to more than one topic, or the group may change · About

Keywords: agents, research, search, web, agent

Key facts. Web, search, and research agents: 1,671 papers on arXiv since 2022 (data through 7 Oct 2026). 193 of its 1,671 mapped papers (12%) cite “WebArena: A Realistic Web Environment for Building Autonomous Agents” (2023, arXiv 2307.13854). 170 of its 1,671 mapped papers (10%) cite “Mind2Web: Towards a Generalist Agent for the Web” (2023, arXiv 2306.06070). 84 new papers 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

84 new papers, 17 Sep – 7 Oct; 14 this week.

  1. This week · 14
  2. From Plausible Hierarchies to Useful Taxonomies: Evaluating Agentic Harnesses on Customer Feedback
  3. Agentic AutoRAG: RAG Pipeline Optimization through Reasoning-Driven Agents
  4. From Delivery to Stateful Exploration: Rethinking the Index for Agentic Search
  5. Leveraging a four-quadrant approach for evaluating Redpine Science
  6. RA-MoWE: Workflow-Affinity Embeddings for Query Clustering and Agentic Workflow Generation
  7. ST-Bench: A Spatial-Temporal Benchmark for Multi-Agent System Generation on Scientific Research Tasks
  8. Stateless Language Agents: Scaling Long-Horizon Automated Research also fits: Software engineering coding agents
  9. Fork-and-Flush: Escaping Idea Basins in Autoresearch Agents
  10. CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling also fits: General LLM agent frameworks and benchmarks
  11. T-Search: An Open Agentic Retriever and Playground for Hard Multi-Step Search
All 84
  1. Programmatic Search Agents: Extending Agentic Search Beyond Query Reformulation
  2. Wikidata Search Traces: A Dataset for Training Knowledge Graph Search Agents
  3. Copies or Sources? Measuring How LLM Aggregators Count Restated Evidence in Multi-Agent Systems also fits: Social simulation and multi-agent debate
  4. Beyond Semantic Similarity: Performance and Costs of Agentic Retrieval for Complex Tasks
  5. Earlier · 70
  6. EPOCH: Reliable Discovery through Evidence-Governed Search
  7. SALUS: Automated Auditing of NL-to-SQL Benchmarks through Weak Supervision of Multi-Agent Output
  8. SearchJev: A Fast and Calibrated System-1 Model for Search Agents
  9. Agentic RAG Evaluation: Budget Allocation Across Questions, Trajectories, and Reads
  10. InvestigationWorlds: An Agentic Environment for Legal Investigation
  11. IdeaScientist: Orchestrating Agents for Grounded Scientific Ideation
  12. HazardWeaver: Scientific Route Selection for Hazard Analysis Agents also fits: Medical and scientific domain agents
  13. MOF-VERIFY: A Failure-Aware Agentic Harness for MOF Hypothesis Verification also fits: Medical and scientific domain agents
  14. Beyond Correctness: Resolving Underspecification in Agentic Text-to-SQL
  15. Open-Endedness Bench: Measuring Epistemic Process from Agent Records
  16. Learning What to Investigate Next: Meta-Reasoning for Long-Horizon Research Agents
  17. MEA: A Reward-Driven Multi-Agent System for Faithful Model Explanations also fits: General LLM agent frameworks and benchmarks
  18. Auditing Web Agent Evaluation on WebArena-Lite: Human Review of Outcomes and Trajectories also fits: Software engineering coding agents
  19. YouRA: A Persistent-State Architecture for Evidence-Traceable Autonomous Research Agents also fits: Medical and scientific domain agents
  20. LawCompass: Navigating from Legal QA to Multi-Agent Deep Research with Grounded Evidence
  21. Breaking Babel: A Self-Evolving Multi-Agent System for Long-Form Subtitle Translation also fits: Agent memory systems
  22. OceanMind: A multi-agent AI system for ocean diagnosis also fits: Medical and scientific domain agents
  23. CompMat-Bench: Benchmarking AI Agents for Computational Materials Science also fits: Medical and scientific domain agents
  24. Legal Research Bench: Measuring End-to-End Reliability in Long-Horizon Legal Research Agents
  25. EurekaBench: Measuring Agentic Ability to Discover New Scientific Insights
  26. How AI Agents Discover Scientific Equations: From Hydrotope Rediscovery to New Water-Wave Amplitudes also fits: Educational agents and math problem solving
  27. Conversational Capture: A Trajectory-Level Framework for Evaluating Generative Engine Optimization in Multi-turn Human-Agent Interaction also fits: Social simulation and multi-agent debate
  28. OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software also fits: Medical and scientific domain agents
  29. RankEvolve: A Reliable Multi-Agent Auto-Research Harness for Evolving Ranking Models
  30. Experimental Experience Modeling for Autonomous Research also fits: Medical and scientific domain agents
  31. DAGent: Evaluate-then-Grow Planning for Deep Research Agents
  32. Search Shapes Conclusions: Auditing Evidence Selection Bias in Deep Research Agents
  33. GraphCert: Bootstrap Agentic Graph Reasoning with Certified Evidence Rubrics
  34. AIM: Agentic Idea Management for Automated Research
  35. Follow the Entities: A Corpus Map for Agentic Search
  36. DeepRewind: Predicting and Repairing Premature Commitments in Deep Research Agents
  37. Reconstructing Implicit Scientific Knowledge: Evaluating LLM Agents through End-to-End Reproduction of Astronomy
  38. From Search to Research: Exploring Search Scaling in Autonomous Quantitative Factor Mining also fits: Financial trading and forecasting agents
  39. MechBench: Can AI Scientific Agents Discover Mechanisms Beyond Phenomenal Laws? also fits: General LLM agent frameworks and benchmarks
  40. SRHarness: A Harness for Agentic Symbolic Regression
  41. From Weak Task Specifications to Scientific Extraction Agents: Optimizing Task Construction also fits: Agent skill learning and evolution
  42. PowerBench: A Benchmark for Agentic Retrieval and Reasoning in Power Systems also fits: General LLM agent frameworks and benchmarks
  43. AgentHop: A Diagnostic Benchmark for Agentic Multi-Hop Scientific Question Answering
  44. FORGE: Form-Optimal Routing of Grounded Evidence for Frozen LLM Agents also fits: General LLM agent frameworks and benchmarks
  45. Agentic High-Dimensional Bayesian Optimization with Hypothesis- and Evidence-Guided Search
  46. TableSeek: Structure-Preserving Agentic Evidence Seeking over Heterogeneous Table Corpora
  47. Evidence-Inference Reconstruction: When The Evidence Is Recalled But The Reasoning Goes Wrong
  48. When Do Agents Help? Embedding, LLM and Agentic Alignment of Classical Texts and Their Translations also fits: General LLM agent frameworks and benchmarks
  49. Robust Hierarchical Structures for Agentic Document Analysis
  50. Overview of the TREC 2025 Million Large Language Models track also fits: General LLM agent frameworks and benchmarks
  51. EmailBench: A Benchmark for Evaluating LLM Agents on Enterprise Email and Productivity Tasks also fits: Tool learning and function calling
  52. Segment-Level Agentic Topic Modeling for Improved Data Exploration and Resource Efficiency
  53. The Hard Part Comes After Search: Benchmarking Web Agents on Synthesizing, Organizing, and Displaying Knowledge also fits: Software engineering coding agents
  54. CARGO: Context-Aware Retrieval-Gated Evaluation of Agentic AI in Production also fits: General LLM agent frameworks and benchmarks
  55. Bridging LLM Agents and Data Spaces: An Architectural Mediation Approach using the Model Context Protocol also fits: Software engineering coding agents
  56. Era by Eon: Benchmarking Enterprise Agents on Hidden Knowledge also fits: Financial trading and forecasting agents
  57. IterSynth: Rethinking Deep Search Agents via Role-Decoupled Iterative Synthesis also fits: Reinforcement learning for LLM agents
  58. ASIRF: An Agentic Framework for Context-Dependent Sensitive Information Redaction also fits: Agent threats, guardrails and governance
  59. RECLAIM: Can Agents Reproduce the Claims of Machine Learning Papers? also fits: Software engineering coding agents
  60. Large Knowledge Model: A Knowledge Foundation for Agentic Science at Scale
  61. Recursive self-improvement of AI research agents also fits: General LLM agent frameworks and benchmarks
  62. Knowledge-as-Skill: A Structural Design for Autonomous Knowledge-Base Use by LLM Agents
  63. When and How Should an Agent Clarify? CIGAsk: Teaching LLMs to Clarify via Counterfactual Information Gain also fits: General LLM agent frameworks and benchmarks
  64. Data Agents: Agentic Data Systems
  65. Monitorable Chart Reasoning Agents via Verifiable Process Rewards also fits: Financial trading and forecasting agents
  66. PrimeSeeker: Capability-Oriented Supervision for Deep Search Agents
  67. Graph Memory for LLM Agents: At What Cost? A Comparative Evaluation of Query, Ingest, and Update Performance Across Graph Database Engines
  68. Bridging Static and Agentic RAG for Taiwanese Historical Question Answering
  69. Semantics Delivery Network: Rethinking Web Retrieval Infrastructure for LLM Agents
  70. Verify, Don't Trust: Agentic Model Development for Video Discovery Retrieval at Scale also fits: Multimodal and embodied agents
  71. The Missing Complement: State-Conditioned Minimal Sufficient Evidence for Coding Agents also fits: Agent memory systems
  72. TRACE: Accountable Agentic Retrieval for Source Discovery in Digital Archives
  73. Point, Revise, Review: Grounded Agentic Analysis in Reactive Notebooks with marimo-lens also fits: Multimodal and embodied agents
  74. AutoData: Agentic Search for Pre-training Data Selection
  75. Self-Evolving Search Index

Foundations

Most cited by papers here

  1. WebArena: A Realistic Web Environment for Building Autonomous Agents2023cited by 193 here
  2. Mind2Web: Towards a Generalist Agent for the Web2023cited by 170 here
  3. WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents2022cited by 116 here
  4. WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models2024cited by 95 here
  5. GPT-4V(ision) is a Generalist Web Agent, if Grounded2024cited by 88 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 181 here
  2. 2023Mind2Web: Towards a Generalist Agent for the Webcited by 170 here
  3. 2023WebArena: A Realistic Web Environment for Building Autonomous Agentscited by 193 here
  4. 2023MetaGPT: Meta Programming for A Multi-Agent Collaborative Frameworkcited by 132 here
  5. 2023AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversationcited by 118 here
  6. 2023A Survey on Large Language Model based Autonomous Agentscited by 121 here
  7. 2023The Rise and Potential of Large Language Model Based Agents: A Surveycited by 61 here
  8. 2024VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Taskscited by 86 here
  9. 2024WebVoyager: Building an End-to-End Web Agent with Large Multimodal Modelscited by 95 here
  10. 2025Agent Laboratory: Using LLM Agents as Research Assistantscited by 74 here
  11. 2025The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Searchcited by 74 here
13 more it builds on, not in this chain
  1. 2022WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agentscited by 116 here
  2. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 179 here
  3. 2023CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Societycited by 62 here
  4. 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 70 here
  5. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 62 here
  6. 2023AgentBench: Evaluating LLMs as Agentscited by 63 here
  7. 2023MLAgentBench: Evaluating Language Agents on Machine Learning Experimentationcited by 78 here
  8. 2024GPT-4V(ision) is a Generalist Web Agent, if Groundedcited by 88 here
  9. 2024SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineeringcited by 81 here
  10. 2024MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineeringcited by 78 here
  11. 2025Search-o1: Agentic Search-Enhanced Large Reasoning Modelscited by 85 here
  12. 2025DeepResearch Bench: A Comprehensive Benchmark for Deep Research Agentscited by 73 here
  13. 2025WebSailor: Navigating Super-human Reasoning for Web Agentcited by 63 here

The 24 most cited of 2,379.

Browse

30 of 1,671, ranked 6–35

  1. 6VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Tasks202486
  2. 7Search-o1: Agentic Search-Enhanced Large Reasoning Models202585
  3. 8MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering202478
  4. 9MLAgentBench: Evaluating Language Agents on Machine Learning Experimentation202378
  5. 10Agent Laboratory: Using LLM Agents as Research Assistants202574
  6. 11The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search202574
  7. 12DeepResearch Bench: A Comprehensive Benchmark for Deep Research Agents202573
  8. 13WebSailor: Navigating Super-human Reasoning for Web Agent202563
  9. 14WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks?202460
  10. 15Agentic Retrieval-Augmented Generation: A Survey on Agentic RAG202558
All 30
  1. 16Data Interpreter: An LLM Agent For Data Science202457
  2. 17ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery202453
  3. 18WebDancer: Towards Autonomous Information Seeking Agency202549
  4. 19An Illusion of Progress? Assessing the Current State of Web Agents202548
  5. 20Beyond Ten Turns: Unlocking Long-Horizon Agentic Search with Large-Scale Asynchronous RL202545
  6. 21InfiAgent-DABench: Evaluating Agents on Data Analysis Tasks202445
  7. 22AgentOccam: A Simple Yet Strong Baseline for LLM-Based Web Agents202442
  8. 23WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learning202442
  9. 24DSBench: How Far Are Data Science Agents from Becoming Data Science Experts?202440
  10. 25The BrowserGym Ecosystem for Web Agent Research202440
  11. 26DS-Agent: Automated Data Science by Empowering Large Language Models with Case-Based Reasoning202439
  12. 27Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers202439
  13. 28BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agent202537
  14. 29AutoWebGLM: A Large Language Model-based Web Navigating Agent202437
  15. 30Deep Research Agents: A Systematic Examination And Roadmap202533
  16. 31WebCanvas: Benchmarking Web Agents in Online Environments202433
  17. 32WebAgent-R1: Training Web Agents via End-to-End Multi-Turn Reinforcement Learning202531
  18. 33WebExplorer: Explore and Evolve for Training Long-Horizon Web Agents202531
  19. 34AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?202428
  20. 35MAC-SQL: A Multi-Agent Collaborative Framework for Text-to-SQL202327