---

Contents

• 🔥 News
• 🤖 End-to-End AI Scientists (35)
• 🔬 Co-Scientists & Research Agents (27)
• 🧰 Open-Source Workbenches (64)
• 🧭 Surveys & Position Papers (38)
• 🔭 Research Stages (72)
• 🌍 Domains (60)
• 🧠 Scientific Foundation Models (14)
• 📊 Benchmarks & Evaluation (85)
• 🏆 Leaderboards & Live Trackers (5)
• 📦 Datasets, Corpora & Environments (55)
• 🛡️ Safety, Ethics & Policy (20)
• 🎓 Tutorials, Blogs & Talks (19)
• 🔗 Related Awesome Lists (3)

---

🔥 News

🧰 2026-09 · Open-Source Workbenches section. Platforms you can actually clone and run now have their own home.

🤗 2026-09 · Hugging Face Daily Papers links. Every entry whose paper has a 🤗 Daily Papers page now carries a direct badge, so you can jump straight to the community discussion and upvotes.

🚀 2026-08 · Repository launch. First release covering AI Scientist systems, co-scientists, benchmarks, datasets, and scientific environments.

💡 Ongoing · PRs welcome. Missing something strong? Open a pull request. One excellent entry beats five weak ones.

---

🤖 End-to-End AI Scientists

Systems that connect several research stages into one loop. Human supervision is still normal, and "autonomous" is not treated as a binary claim.

• OmniScientist, "An Omni-Modal Omni-Discipline AI Scientist".
• The AI Scientist, "Towards Fully Automated Open-Ended Scientific Discovery".
• The AI Scientist-v2, "Workshop-Level Automated Scientific Discovery via Agentic Tree Search".
• Robin, "A multi-agent system for automating scientific discovery".
• data-to-paper, "Autonomous LLM-driven research from data to human-verifiable research papers".
• Agent Laboratory, "Using LLM Agents as Research Assistants".
• AutoResearchClaw, "Self-Reinforcing Autonomous Research with Human-AI Collaboration".
• EvoScientist, "Towards Multi-Agent Evolving AI Scientists for End-to-End Scientific Discovery".
• DeepScientist, "Advancing Frontier-Pushing Scientific Findings Progressively".
• Kosmos, "An AI Scientist for Autonomous Discovery".
• Denario, "The Denario project: Deep knowledge AI agents for scientific discovery".
• CycleResearcher, "Improving Automated Research via Automated Review".
• Dolphin, "Moving Towards Closed-loop Auto-research through Thinking, Practice, and Feedback".
• CORAL, "Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery".
• ASI-Evolve, "AI Accelerates AI".
• Darwin Godel Machine, "Open-Ended Evolution of Self-Improving Agents".
• AlphaEvolve, "A coding agent for scientific and algorithmic discovery".
• FunSearch, "Mathematical discoveries from program search with large language models".
• ShinkaEvolve, "Towards Open-Ended And Sample-Efficient Program Evolution".
• MLE-STAR, "Machine Learning Engineering Agent via Search and Targeted Refinement".
• ML-Master 2.0, "Toward Ultra-Long-Horizon Agentic Science: Cognitive Accumulation for Machine Learning Engineering".
• Kolb-Based Experiential Learning (Agent K), "Kolb-Based Experiential Learning for Generalist Agents with Human-Level Kaggle Data Science Performance".
• AutoKaggle, "A Multi-Agent Framework for Autonomous Data Science Competitions".
• DS-Agent, "Automated Data Science by Empowering Large Language Models with Case-Based Reasoning".
• AutoML-Agent, "A Multi-Agent LLM Framework for Full-Pipeline AutoML".
• MLR-Copilot, "Autonomous Machine Learning Research based on Large Language Models Agents".
• AI-Newton, "A Concept-Driven Physical Law Discovery System without Prior Physical Knowledge".
• AtomAgents, "Alloy design and discovery through physics-aware multi-modal multi-agent artificial intelligence".
• MatAgent, "Accelerated Inorganic Materials Design with Generative AI Agents".
• El Agente, "An Autonomous Agent for Quantum Chemistry".
• LabOS, "The AI-XR Co-Scientist That Sees and Works With Humans".
• ORGANA, "A Robotic Assistant for Automated Chemistry Experimentation and Characterization".
• ProtAgents, "Protein discovery via large language model multi-agent collaborations combining physics and machine learning".
• STELLA, "Self-Evolving LLM Agent for Biomedical Research".
• CRISPR-GPT, "CRISPR-GPT for Agentic Automation of Gene-editing Experiments".

---

🔬 Co-Scientists & Research Agents

Highly relevant to AI Scientist research, but focused on part of the loop or explicitly keeping the human scientist as decision maker.

• Co-Scientist, "Accelerating scientific discovery with Co-Scientist".
• PaperQA2, "Language agents achieve superhuman synthesis of scientific knowledge".
• OpenScholar, "Synthesizing Scientific Literature with Retrieval-augmented LMs".
• Biomni, "A General-Purpose Biomedical AI Agent".
• Coscientist, "Autonomous chemical research with large language models".
• ChemCrow, "Augmenting large-language models with chemistry tools".
• ResearchAgent, "Iterative Research Idea Generation over Scientific Literature with Large Language Models".
• Aviary, "training language agents on challenging scientific tasks".
• SciAgents, "Automating scientific discovery through multi-agent intelligent graph reasoning".
• ToolUniverse, "An open platform for democratizing AI scientists".
• CoI-Agent (Chain of Ideas), "Chain of Ideas: Revolutionizing Research Via Novel Idea Development with LLM Agents".
• SciMaster / X-Master, "SciMaster: Towards General-Purpose Scientific AI Agents, Part I. X-Master as Foundation: Can We Lead on Humanity's Last Exam?".
• BioDiscoveryAgent, "An AI Agent for Designing Genetic Perturbation Experiments".
• GeneAgent, "Self-verification Language Agent for Gene Set Knowledge Discovery using Domain Databases".
• TAIS, "Toward a Team of AI-made Scientists for Scientific Discovery from Gene Expression Data".
• GenoMAS, "A Multi-Agent Framework for Scientific Discovery via Code-Driven Gene Expression Analysis".
• dZiner, "Rational Inverse Design of Materials with AI Agents".
• ChemAgent, "Self-updating Library in Large Language Models Improves Chemical Reasoning".
• LLMatDesign, "Autonomous Materials Discovery with Large Language Models".
• BioResearcher, Multi-agent pipeline searches literature and datasets, then drafts and reviews dry and wet-lab biomedical experimental protocols.
• AstroAgents, "A Multi-Agent AI for Hypothesis Generation from Mass Spectrometry Data".
• The AI Cosmologist, "The AI Cosmologist I: An Agentic System for Automated Data Analysis".
• Nova, "An Iterative Planning and Search Approach to Enhance Novelty and Diversity of LLM Generated Ideas".
• SciSciGPT, "Advancing Human-AI Collaboration in the Science of Science".
• SciAgent, "Tool-augmented Language Models for Scientific Reasoning".
• Elicit, Screens and extracts structured data from 125M papers, automating systematic-review screening into reusable tables.
• Undermind, Iteratively reads and scores hundreds of papers and follows citation trails until a search is exhaustive.

---

🧰 Open-Source Workbenches

Platforms you can clone, install, and drive today.

🖥️ End-to-end research platforms

• OmniScientist, Omni-modal, omni-discipline AI scientist you can run locally across heterogeneous scientific evidence.
• Open Science Desktop, Local-first desktop workbench wiring agents, notebooks, runs, figures and review into one auditable provenance trail.
• InternAgent, "When Agent Becomes the Scientist -- Building Closed-Loop System from Hypothesis to Verification".
• RD-Agent, "R&D-Agent: An LLM-Agent Framework Towards Autonomous Data Science".
• freephdlabor, "Build Your Personalized Research Group: A Multiagent Framework for Continual and Interactive Science Automation".
• The Virtual Lab, "The Virtual Lab of AI agents designs new SARS-CoV-2 nanobodies".
• Curie, "Toward Rigorous and Automated Scientific Experimentation with AI Agents".
• TxAgent, "An AI Agent for Therapeutic Reasoning Across a Universe of Tools".
• Galaxy, "Galaxy for accessible, reproducible, and collaborative data analyses: 2026 update".
• MLE-Agent, Plans and implements ML engineering work with arXiv integration and code retrieval.

📖 Literature and deep-research platforms

• STORM and Co-STORM, "Assisting in Writing Wikipedia-like Articles From Scratch with Large Language Models".
• Asta, AI2's science agent family, reproducible and benchmarkable against a rigorous multi-task research suite.
• GPT Researcher, Autonomous deep research over web and local documents with any provider, emitting a cited report.
• DeerFlow, Long-horizon agent harness that researches, writes code, and produces artifacts.
• Local Deep Research, Fully local, encrypted research agent over arXiv, PubMed, and your own private document collection.
• OpenResearcher, "Unleashing AI for Accelerated Scientific Research".
• DeepResearchAgent, Hierarchical planner plus specialist agents for deep research and general task execution.
• DeepLiterature, Open research assistant combining search, code execution, link resolution and information expansion.
• Deep Research from Scratch, LangGraph reference implementation of a deep-research agent, the maintained successor to Open Deep Research.

🔬 Lab automation and self-driving labs

• Opentrons, Write Python protocols and execute them on physical Flex and OT-2 liquid-handling robots.
• PyLabRobot, "An open-source, hardware-agnostic interface for liquid-handling robots and accessories".
• Bluesky, Orchestrates beamline and laboratory experiments plus data acquisition, in production at NSLS-II.
• Atlas, "a brain for self-driving laboratories".
• Olympus, "a benchmarking framework for noisy optimization and experiment planning".
• Self-Driving Lab Demo, Build and run a real low-cost autonomous experimentation rig from dimmable LEDs and a spectrophotometer.

🧩 Libraries, skill packs and components

• Scientific Agent Skills, 165 validated science skills plus database connectors, installable into Claude Code, Cursor or Codex.
• AI4S Skills, Agent skills for topic exploration, literature survey, experiments, paper writing and integrity audit.
• AIDE, "AI-Driven Exploration in the Space of Code".
• TinyScientist, "An Interactive, Extensible, and Controllable Framework for Building Research Agents".
• ResearchTown, "Simulator of Human Research Community".
• Virtual Scientists, Multi-agent simulation of science-of-science dynamics over real publication data.
• Intern-S1, "A Scientific Multimodal Foundation Model".
• Data Formulator, "Data Formulator 2: Iterative Creation of Data Visualizations, with AI Transforming Data Along the Way".
• OpenChemIE, "An Information Extraction Toolkit for Chemistry Literature".
• MolScribe, "Robust Molecular Structure Recognition with Image-to-Graph Generation".

🧱 General agent frameworks these are built on

• MetaGPT and Data Interpreter, "Data Interpreter: An LLM Agent For Data Science".
• CAMEL, "Communicative Agents for "Mind" Exploration of Large Language Model Society".
• OWL, "Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation".
• ChatDev, "Communicative Agents for Software Development".
• smolagents, Minimal library for code-writing agents, shipping the reference Open Deep Research implementation.

🔧 More runnable systems

• autoresearch (karpathy), Gives an agent a real single-GPU nanochat training setup and lets it modify code, train, evaluate, keep or discard.
• ARIS (Auto-Research-In-Sleep), Markdown-only skill pack for autonomous ML research providing cross-model review loops, idea discovery and experiment automation.
• Paper2Agent, "Reimagining Research Papers As Interactive and Reliable AI Agents".
• Paper2Code, "Automating Code Generation from Scientific Papers in Machine Learning".
• claude-scholar, Semi-automated research assistant spanning ideation, coding, experiments, writing and publication across Claude Code, Codex, Kimi and OpenCode.
• Dr. Claw, "An AI Scientist Workspace for Vibe Research".
• ScienceClaw, Self-evolving research colleague with 285 skills across 28 disciplines and persistent memory over literature and databases.
• MLGym, "A New Framework and Benchmark for Advancing AI Research Agents".
• AIRA / aira-dojo, "AI Research Agents for Machine Learning: Search, Exploration, and Generalization in MLE-bench".
• DeepReview / DeepReviewer, "DeepReview: Improving LLM-based Paper Review with Human-like Deep Thinking Process".
• CMBAgent, "Open Source Planning & Control System with Language Agents for Autonomous Scientific Discovery".
• LLM-SR, "Scientific Equation Discovery via Programming with Large Language Models".
• AI Hilbert, "Evolving scientific discovery by unifying data and background knowledge with AI Hilbert".
• PaSa, "An LLM Agent for Comprehensive Academic Paper Search".
• AutoSurvey, "Large Language Models Can Automatically Write Surveys".
• AutoSota, CLI and leaderboard that autonomously optimizes existing research codebases, publishing results only when an internal ledger confirms improvement.
• ChemGraph, "An Agentic Framework for Computational Chemistry Workflows".
• MDCrow, "Automating Molecular Dynamics Workflows with Large Language Models".
• LLaMP, "Large Language Model Made Powerful for High-fidelity Materials Knowledge Retrieval and Distillation".
• HoneyComb, "A Flexible LLM-Based Agent System for Materials Science".
• CellAgent, "An LLM-driven Multi-Agent Framework for Automated Single-cell Data Analysis".
• ResearchCodeAgent, "An LLM Multi-Agent System for Automated Codification of Research Methodologies".
• EvoMaster, Foundational evolving-agent framework reimplementing the SciMaster line including ML-Master, X-Master and Browse-Master.
• LabClaw, Skills-only operating layer for LabOS, with no engine of its own.

⚠️ License traps worth knowing

AI-Researcher (HKUDS/AI-Researcher, 5.7k stars) ships substantial code but has no license file at all, so all rights are reserved. Coscientist (gomesgroup/coscientist) carries a Commons Clause rider forbidding sale, which is not OSI open source. The AI Scientist and v2 use a custom AI Scientist Source Code License derived from the Responsible AI license, carrying use restrictions. UniScientist (UniPat-AI/UniScientist) has no license file.

🧭 Surveys & Position Papers

• From AI for Science to Agentic Science, "A Survey on Autonomous Scientific Discovery".
• Agentic AI for Scientific Discovery, "A Survey of Progress, Challenges, and Future Directions".
• From Automation to Autonomy, "A Survey on Large Language Models in Scientific Discovery".
• Agent Systems for Academic Research Automation, Survey of research agents across literature, ideation, experimentation, and paper production.
• Exploring the role of LLMs in the scientific method, "Exploring the role of large language models in the scientific method: from hypothesis to discovery".
• Automated Scientific Discovery, "From Equation Discovery to Autonomous Discovery Systems".
• AI4Research, "A Survey of Artificial Intelligence for Scientific Research".
• LLM4SR, "A Survey on Large Language Models for Scientific Research".
• A Comprehensive Survey of Scientific Large Language Models and Their Applications in Scientific Discovery
• A Survey of Scientific Large Language Models, "From Data Foundations to Agent Frontiers".
• Towards Scientific Intelligence, "A Survey of LLM-based Scientific Agents".
• Transforming Science with Large Language Models, "A Survey on AI-assisted Scientific Discovery, Experimentation, Content Generation, and Evaluation".
• From Hypothesis to Publication, "A Comprehensive Survey of AI-Driven Research Support Systems".
• A Survey of AI Scientists
• Autonomous Research Agents, "A Survey of AI Scientists and the Verification Gap".
• Deep Research, "A Survey of Autonomous Research Agents".
• A Comprehensive Survey of Deep Research, "Systems, Methodologies, and Applications".
• A Survey on Hypothesis Generation for Scientific Discovery in the Era of Large Language Models
• Large Language Models for Scientific Idea Generation, "A Creativity-Centered Survey".
• A review of large language models and autonomous agents in chemistry
• A Survey on Large Language Model-based Agents for Statistics and Data Science
• Measuring Data Science Automation, "A Survey of Evaluation Tools for AI Assistants and Agents".
• LLM-Based Data Science Agents, "A Survey of Capabilities, Challenges, and Future Directions".
• The rise of self-driving labs in chemical and materials sciences
• Empowering biomedical discovery with AI agents
• Nobel Turing Challenge, "creating the engine for scientific discovery".
• Exploring the use of AI authors and reviewers at Agents4Science
• Evolving Roles of LLMs in Scientific Innovation, "Assistant, Collaborator, Scientist, and Evaluator".
• Position, "AI Agents in Scientific Teams Should Be Studied as Human-Agent Systems".
• Scaling Laws in Scientific Discovery with AI and Robot Scientists
• The Ideation-Execution Gap, "Execution Outcomes of LLM-Generated versus Human Research Ideas".
• Can AI make scientific discoveries?
• The More You Automate, the Less You See, "Hidden Pitfalls of AI Scientist Systems".
• Why LLMs Aren't Scientists Yet, "Lessons from Four Autonomous Research Attempts".
• AI Research Agents Narrow Scientific Exploration
• Can AI agents conduct open-ended AI research? Early evidence from two case studies
• How Far Are AI Scientists from Changing the World?
• Fully Autonomous AI Agents Should Not be Developed

---

🔭 Research Stages

Work that targets one stage of the research loop rather than the whole thing.

💡 Ideation & hypothesis generation

• Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers
• SciMON, "Scientific Inspiration Machines Optimized for Novelty".
• SciPIP, "An LLM-based Scientific Paper Idea Proposer".
• MOOSE-Chem, "Large Language Models for Rediscovering Unseen Chemistry Scientific Hypotheses".
• Large Language Models for Automated Open-domain Scientific Hypotheses Discovery
• MOOSE-Chem3, "Toward Experiment-Guided Hypothesis Ranking via Simulated Experimental Feedback".
• Many Heads Are Better Than One, "Improved Scientific Idea Generation by A LLM-Based Multi-Agent System".
• IRIS, "Interactive Research Ideation System for Accelerating Scientific Discovery".
• HypoBench, "Towards Systematic and Principled Benchmarking for Hypothesis Generation".
• Literature Meets Data, "A Synergistic Approach to Hypothesis Generation".
• AI Idea Bench 2025, "AI Research Idea Generation Benchmark".
• IdeaBench, "Benchmarking Large Language Models for Research Idea Generation".
• Can Large Language Models Unlock Novel Scientific Research Ideas?
• Sparks of Science, "Hypothesis Generation Using Structured Paper Data".
• Large Language Models as Biomedical Hypothesis Generators, "A Comprehensive Evaluation".
• Spark, "A System for Scientifically Creative Idea Generation".
• LDC: Learning to Generate Research Idea with Dynamic Control
• TrustResearcher, "Automating Knowledge-Grounded and Transparent Research Ideation with Multi-Agent Collaboration".
• Is this Idea Novel? An Automated Benchmark for Judgment of Research Ideas

📖 Literature review & retrieval

• SurveyForge, "On the Outline Heuristics, Memory-Driven Generation, and Multi-dimensional Evaluation for Automated Survey Writing".
• SurveyX, "Academic Survey Automation via Large Language Models".
• LitSearch, "A Retrieval Benchmark for Scientific Literature Search".
• ArxivDIGESTables, "Synthesizing Scientific Literature into Tables using Language Models".
• LitLLM, "A Toolkit for Scientific Literature Review".
• SPAR, "Scholar Paper Retrieval with LLM-based Agents for Enhanced Academic Search".
• Citegeist, "Automated Generation of Related Work Analysis on the arXiv Corpus".
• SurveyBench, "Can LLM(-Agents) Write Academic Surveys that Align with Reader Needs?".
• SurveyGen, "Quality-Aware Scientific Survey Generation with Large Language Models".
• ScholarGym, "Benchmarking Large Language Model Capabilities in the Information-Gathering Stage of Deep Research".
• Patience is all you need! An agentic system for performing scientific literature review
• Multi-Turn Agentic Scientific Literature Search via Workflow Induction

🧪 Experimentation & code execution

• AutoMind, "Adaptive Knowledgeable Agent for Automated Data Science".
• A Self-Improving Coding Agent
• CodeScientist, "End-to-End Semi-Automated Scientific Discovery with Code-based Experimentation".
• MLZero, "A Multi-Agent System for End-to-end Machine Learning Automation".
• AutoReproduce, "Automatic AI Experiment Reproduction with Paper Lineage".
• I-MCTS, "Enhancing Agentic AutoML via Introspective Monte Carlo Tree Search".
• MLAgentBench, "Evaluating Language Agents on Machine Learning Experimentation".
• EXP-Bench, "Can AI Conduct AI Research Experiments?".

✍️ Scientific writing

• AutoP2C, "An LLM-Based Agent Framework for Code Repository Generation from Multimodal Content in Academic Papers".
• Paper2Poster, "Towards Multimodal Poster Automation from Scientific Papers".
• P2P: Automated Paper-to-Poster Generation and Fine-Grained Benchmark
• AutoPresent, "Designing Structured Visuals from Scratch".
• Paper2Video, "Automatic Video Generation from Scientific Papers".
• ScholarCopilot, "Training Large Language Models for Academic Writing with Accurate Citations".
• SciLitLLM, "How to Adapt LLMs for Scientific Literature Understanding".
• ScholaWrite, "A Dataset of End-to-End Scholarly Writing Process".
• PaperOrchestra, "A Multi-Agent Framework for Automated AI Research Paper Writing".
• DeTikZify, "Synthesizing Graphics Programs for Scientific Figures and Sketches with TikZ".
• AutomaTikZ, "Text-Guided Synthesis of Scientific Vector Graphics with TikZ".
• SciDoc2Diagrammer-MAF, "Towards Generation of Scientific Diagrams from Documents guided by Multi-Aspect Feedback Refinement".
• Attribution in Scientific Literature, "New Benchmark and Methods".
• Do Language Models Know When They're Hallucinating References?
• From Sparse to Dense, "GPT-4 Summarization with Chain of Density Prompting".

🧾 Peer review & verification

• Can large language models provide useful feedback on research papers? A large-scale empirical analysis
• OpenReviewer, "A Specialized Large Language Model for Generating Critical Scientific Paper Reviews".
• AgentReview, "Exploring Peer Review Dynamics with LLM Agents".
• Reviewer2, "Optimizing Review Generation Through Prompt Generation".
• LLMs Assist NLP Researchers, "Critique Paper (Meta-)Reviewing".
• Monitoring AI-Modified Content at Scale, "A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews".
• Is Your Paper Being Reviewed by an LLM? Benchmarking AI Text Detection in Peer Review
• BadScientist, "Can a Research Agent Write Convincing but Unsound Papers that Fool LLM Reviewers?".
• When AI Co-Scientists Fail, "SPOT-a Benchmark for Automated Verification of Scientific Research".
• AI-Assisted Peer Review at Scale, "The AAAI-26 AI Review Pilot".
• Gaming AI-Assisted Peer Reviews Poses New Risks to the Scientific Community
• Unveiling the Merits and Defects of LLMs in Automatic Review Generation for Scientific Papers
• Fact or Fiction, "Verifying Scientific Claims".
• MultiVerS, "Improving scientific claim verification with weak supervision and full-document context".
• SciClaimHunt, "A Large Dataset for Evidence-based Scientific Claim Verification".

🧠 Knowledge, tools & environments

• KARMA, "Leveraging Multi-Agent LLMs for Automated Knowledge Graph Enrichment".
• Agentic deep graph reasoning yields self-organizing knowledge networks
• Experiences with Model Context Protocol Servers for Science and High Performance Computing

---

---

🌍 Domains

Systems built for one science rather than for research in general.

⚗️ Chemistry & materials

• A-Lab / AlabOS, "AlabOS: A Python-based Reconfigurable Workflow Management Framework for Autonomous Laboratories".
• GNoME, "Scaling deep learning for materials discovery".
• MatterGen, "a generative model for inorganic materials design".
• MatterSim, "A Deep Learning Atomistic Model Across Elements, Temperatures and Pressures".
• ChemOS, "An orchestration software to democratize autonomous discovery".
• ChemOS 2.0, "An orchestration architecture for chemical self-driving laboratories".
• Chemist-X, "Large Language Model-empowered Agent for Reaction Condition Recommendation in Chemical Synthesis".
• ChemReasoner, "Heuristic Search over a Large Language Model's Knowledge Space using Quantum-Chemical Feedback".
• CACTUS, "Chemistry Agent Connecting Tool-Usage to Science".
• SynAsk, "Unleashing the Power of Large Language Models in Organic Synthesis".
• AlphaFlow, "autonomous discovery and optimization of multi-step chemistry using a self-driven fluidic lab guided by reinforcement learning".
• RoboChem, "Automated self-optimization, intensification, and scale-up of photocatalysis in flow".
• LLM-RDF, "An automatic end-to-end chemical synthesis development platform powered by large language models".
• Mobile Robotic Chemist, "A mobile robotic chemist".
• Autonomous Mobile Robots for Exploratory Synthetic Chemistry
• Multiagent-Driven Robotic AI Chemist, "A Multiagent-Driven Robotic AI Chemist Enabling Autonomous Chemical Research On Demand".
• The World Avatar, "A dynamic knowledge graph approach to distributed self-driving laboratories".

🧬 Biology & medicine

• AMIE, "Towards Conversational Diagnostic AI".
• PathChat, "A multimodal generative AI copilot for human pathology".
• SpatialAgent, "An autonomous AI agent for spatial biology".
• DrugAgent, "Automating AI-aided Drug Discovery Programming through LLM Multi-Agent Collaboration".
• BioPlanner, "Automatic Evaluation of LLMs on Protocol Planning in Biology".
• BioMARS, "A Multi-Agent Robotic System for Autonomous Biological Experiments".
• MedRAX, "Medical Reasoning Agent for Chest X-ray".

🔭 Physics, astronomy & earth

• AI Feynman, "a Physics-Inspired Method for Symbolic Regression".
• PySR, "Interpretable Machine Learning for Science with PySR and SymbolicRegression.jl".
• AlphaTensor, "Discovering faster matrix multiplication algorithms with reinforcement learning".
• DeepMind Tokamak Plasma Control, "Magnetic control of tokamak plasmas through deep reinforcement learning".
• Scientific Generative Agent, "LLM and Simulation as Bilevel Optimizers: A New Paradigm to Advance Physical Scientific Discovery".
• KAN, "Kolmogorov-Arnold Networks".
• ClimateGPT, "Towards AI Synthesizing Interdisciplinary Research on Climate Change".
• GeoGalactica, "A Scientific Large Language Model in Geoscience".

➗ Mathematics

• AlphaProof, "Olympiad-level formal mathematical reasoning with reinforcement learning".
• AlphaGeometry, "Solving olympiad geometry without human demonstrations".
• AlphaGeometry2, "Gold-medalist Performance in Solving Olympiad Geometry with AlphaGeometry2".
• LeanDojo, "Theorem Proving with Retrieval-Augmented Language Models".
• Lean Copilot, "Large Language Models as Copilots for Theorem Proving in Lean".
• LeanAgent, "Lifelong Learning for Formal Theorem Proving".
• DeepSeek-Prover, "Advancing Theorem Proving in LLMs through Large-Scale Synthetic Data".
• DeepSeek-Prover-V1.5, "Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search".
• DeepSeek-Prover-V2, "Advancing Formal Mathematical Reasoning via Reinforcement Learning for Subgoal Decomposition".
• Kimina-Prover, "Kimina-Prover Preview: Towards Large Formal Reasoning Models with Reinforcement Learning".
• Seed-Prover, "Solving Formal Math Problems by Decomposition and Iterative Reflection".
• Goedel-Prover, "A Frontier Model for Open-Source Automated Theorem Proving".
• Goedel-Prover-V2, Scaffolded data synthesis plus verifier-guided self-correction, shipping open 8B and 32B theorem-proving weights.
• NuminaMath / AIMO Progress Prize, Winning AI Mathematical Olympiad pipeline plus the 860k chain-of-thought corpus that became the standard open math dataset.
• PutnamBench, "Evaluating Neural Theorem-Provers on the Putnam Mathematical Competition".
• miniF2F, "a cross-system benchmark for formal Olympiad-level mathematics".
• InternLM-Math, "Open Math Large Language Models Toward Verifiable Reasoning".
• Equational Theories Project, Tao-led crowdsourced Lean formalization settling all 22 million implications between 4694 magma equational laws.
• Formal Conjectures, "An Open and Evolving Benchmark for Verified Discovery in Mathematics".
• Aristotle, "IMO-level Automated Theorem Proving".

🏛️ Social science & humanities

• Generative Agents, "Interactive Simulacra of Human Behavior".
• Automated Social Science, "Language Models as Scientist and Subjects".
• LLM Agents Grounded in Self-Reports, "LLM Agents Grounded in Self-Reports Enable General-Purpose Simulation of Individuals".
• EconAgent, "Large Language Model-Empowered Agents for Simulating Macroeconomic Activities".
• AgentSociety, "Large-Scale Simulation of LLM-Driven Generative Agents Advances Understanding of Human Behaviors and Society".
• OASIS, "Open Agent Social Interaction Simulations with One Million Agents".
• GenSim, General social simulation platform with scripted and free-form modes, supporting up to 100k agents with error correction.
• Can Large Language Models Transform Computational Social Science?

---

🧠 Scientific Foundation Models

Open or landmark models that autonomous discovery systems call as tools.

• AlphaFold 3, "Accurate structure prediction of biomolecular interactions with AlphaFold 3".
• ESM3, "Simulating 500 million years of evolution with a language model".
• Evo 2, "Genome modeling and design across all domains of life with Evo 2".
• Boltz-2, "Towards Accurate and Efficient Binding Affinity Prediction".
• Chai-1, "Decoding the molecular interactions of life".
• Aurora, "A Foundation Model for the Earth System".
• AstroLLaMA, "Towards Specialized Foundation Models in Astronomy".
• Galactica, "A Large Language Model for Science".
• SciBERT, "A Pretrained Language Model for Scientific Text".
• ChemDFM, "Developing ChemDFM as a large language foundation model for chemistry".
• nach0, "Multimodal Natural and Chemical Languages Foundation Model".
• DPLM, "Diffusion Language Models Are Versatile Protein Learners".
• GenomeOcean, Efficient 4B genome foundation model trained on 600 Gbp of metagenomic assemblies for de novo sequence generation.
• scGPT, "toward building a foundation model for single-cell multi-omics using generative AI".

---

📊 Benchmarks & Evaluation

Evaluation is the bottleneck. These test scientific knowledge, literature grounding, code and experiment execution, whole research workflows, or interaction with scientific environments.

• AstaBench, "Rigorous Benchmarking of AI Agents with a Scientific Research Suite".
• ScienceBoard, "Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows".
• ScienceAgentBench, "Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery".
• PaperBench, "Evaluating AI's Ability to Replicate AI Research".
• CORE-Bench, "Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark".
• SciCode, "A Research Coding Benchmark Curated by Scientists".
• MLE-bench, "Evaluating Machine Learning Agents on Machine Learning Engineering".
• LAB-Bench, "Measuring Capabilities of Language Models for Biology Research".
• AIRS-Bench, "a Suite of Tasks for Frontier AI Research Science Agents".

🤖 Research automation

• RE-Bench, "Evaluating frontier AI R&D capabilities of language model agents against human experts".
• MLRC-Bench, "Can Language Agents Solve Machine Learning Research Challenges?".
• MLR-Bench, "Evaluating AI Agents on Open-Ended Machine Learning Research".
• InnovatorBench, "Evaluating Agents' Ability to Conduct Innovative LLM Research".
• MLE-Dojo, "Interactive Environments for Empowering LLM Agents in Machine Learning Engineering".
• AutoSDT, "Scaling Data-Driven Discovery Tasks Toward Open Co-Scientists".
• The Automated LLM Speedrunning Benchmark, "Reproducing NanoGPT Improvements".
• ResearchGym, "Evaluating Language Model Agents on Real-World AI Research".
• MLS-Bench, "A Holistic and Rigorous Assessment of AI Systems on Building Better AI".
• FML-bench, "Benchmarking Machine Learning Agents for Scientific Research".
• NatureBench, "Can Coding Agents Match the Published SOTA of Nature-Family Papers?".
• FIRE-Bench, "Evaluating AI Agents on the Rediscovery of Scientific Insights".
• SGI-Bench, "Probing Scientific General Intelligence of LLMs with Scientist-Aligned Workflows".
• DeltaML-Bench, "Evaluating Machine Learning Agents on Real-World Research Repositories".
• AARRI-Bench, "Act As a Real Researcher: A Suite of Benchmarks Evaluating Frontier LLMs and Agentic Harnesses in Research Lifecycle".
• SciAgentArena, "Benchmarking AI Agents for Addressing Scientific Challenges Across Scales".

💻 Scientific coding

• SUPER, "Evaluating Agents on Setting Up and Executing Tasks from Research Repositories".
• DSBench, "How Far Are Data Science Agents from Becoming Data Science Experts?".
• DABStep, "Data Agent Benchmark for Multi-step Reasoning".
• BixBench, "a Comprehensive Benchmark for LLM-b