Autonomous chemical research with large language models
2023/12/20 by Daniil A. Boiko, Robert MacKnight, Ben Kline +1 · 174 citations
Engineering · Materials Science · #Ferroelectric and Negative Capacitance Devices #Innovative Microfluidic and Catalytic Techniques Innovation #Machine Learning in Materials Science
paper · pdf · doi:10.1038/s41586-023-06792-0
openalex publication_date 2023/12/20 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/03
Abstract
Abstract Transformer-based large language models are making significant strides in various fields, such as natural language processing 1–5 , biology 6,7 , chemistry 8–10 and computer programming 11,12 . Here, we show the development and capabilities of Coscientist, an artificial intelligence system driven by GPT-4 that autonomously designs, plans and performs complex experiments by incorporating large language models empowered by tools such as internet and documentation search, code execution and experimental automation. Coscientist showcases its potential for accelerating research across six diverse tasks, including the successful reaction optimization of palladium-catalysed cross-couplings, while exhibiting advanced capabilities for (semi-)autonomous experimental design and execution. Our findings demonstrate the versatility, efficacy and explainability of artificial intelligence systems like Coscientist in advancing research.
Citations
Cited by
- SciMON: Scientific Inspiration Machines Optimized for Novelty
- Using a GPT-5-driven autonomous lab to optimize the cost and titer of cell-free protein synthesis
- Learning from Wildfire Decision Support: large language model analysis of barriers to fire spread in a census of large wildfires in the United States (2011–2023)
- The missing data for intelligent scientific instruments
- Predicting Reactivity with Machine Learning
- Self-driving laboratories in Japan
- Report of the 2025 Workshop on Next-Generation Ecosystems for Scientific Computing: Harnessing Community, Software, and AI for Cross-Disciplinary Team Science
- Can Current Agents Close the Discovery-to-Application Gap? A Case Study in Minecraft
- Atom-anchored LLMs speak Chemistry: A Retrosynthesis Demonstration
- Synergizing chemical and AI communities for advancing laboratories of the future
- Guidelines needed for the use of AI in the preparation or review of IRB, IBC, and IACUC applications
- Larger and more instructable language models become less reliable
- A Survey of AI Scientists
- Towards end-to-end automation of AI research
- Shoot First, Ask Questions Later? Building Rational Agents that Explore and Act Like People
- ComProScanner: A multi-agent based framework for composition-property structured data extraction from scientific literature
- Cross-disciplinary perspectives on the potential for artificial intelligence across chemistry
- The role of large language models in self-care: a study and benchmark on medicines and supplement guidance accuracy
- Foundation Models for Scientific Discovery: From Paradigm Enhancement to Paradigm Transition
- Multi-Agent Design Assistant for the Simulation of Inertial Fusion Energy
- Generative AI for Biosciences: Emerging Threats and Roadmap to Biosecurity
- An agentic artificially intelligent X-ray scientist
- The Last Human-Written Paper: Agent-Native Research Artifacts
- AI Agents Can Already Autonomously Perform Experimental High Energy Physics
- Autonomous Agents for Scientific Discovery: Orchestrating Scientists, Language, Code, and Physics
- ConDABench: Interactive Evaluation of Language Models for Data Analysis
- Artificial intelligence tools expand scientists’ impact but contract science’s focus
- Multi-agentic AI framework for end-to-end atomistic simulations
- Scientific Algorithm Discovery by Augmenting AlphaEvolve with Deep Research
- RareAgent: Self-Evolving Reasoning for Drug Repurposing in Rare Diseases
- Large Language Models Achieve Gold Medal Performance at the International Olympiad on Astronomy & Astrophysics (IOAA)
- Multimodal AI agents for capturing and sharing laboratory practice
- Zephyrus: An Agentic Framework for Weather Science
- Speak to a Protein: An Interactive Multimodal Co-Scientist for Protein Analysis
- Posthuman Cartography? Rethinking Artificial Intelligence, Cartographic Practices, and Reflexivity
- Large Language Models Inference Engines based on Spiking Neural Networks
- LLM Agents for Knowledge Discovery in Atomic Layer Processing
- AutoLabs: Cognitive Multi-Agent Systems with Self-Correction for Autonomous Chemical Experimentation
- Agentic Exploration of Physics Models
- Agentic Services Computing
- Dehumanizing the human, humanizing the machine: organic consciousness as a hallmark of the persistence of the human against the backdrop of artificial intelligence
- From What to Why: A Multi-Agent System for Evidence-based Chemical Reaction Condition Reasoning
- Democratizing AI scientists using ToolUniverse
- Artificial intelligence in drug development
- Bridging the Gap Between Scientific Laws Derived by AI Systems and Canonical Knowledge via Abductive Inference with AI-Noether
- AI for Sustainable Future Foods
- AOT*: Efficient Synthesis Planning via LLM-Empowered AND-OR Tree Search
- Safety-Gated Agentic Supervisory Control on a Coupled Distillation Benchmark: Regime Map, Auditable Gate, and Co-Design Findings
- LabEvolver: Training-Free Experience Evolution for Safe and Grounded Wet-Lab Agents
- SciSchema.org: A Multidisciplinary Collection of Schemas for Structured Scientific Process Descriptions
- Can AI Follow In Einstein's Footsteps?
- Adsorb-Agent: autonomous identification of stable adsorption configurations via a large language model agent
- QCalEval: Benchmarking Vision-Language Models for Quantum Calibration Plot Understanding
- Optimizing ground state preparation protocols with autoresearch
- The Evolving Role of Programming and LLMs in the Development of Self-Driving Laboratories
- Agentic Artificial Intelligence for Multistage Physics Experiments at a Large-Scale User Facility Particle Accelerator
- ChemOrch: Empowering LLMs with Chemical Intelligence via Synthetic Instructions
- CFDLLMBench: A Benchmark Suite for Evaluating Large Language Models in Computational Fluid Dynamics
- From Language to Action: A Review of Large Language Models as Autonomous Agents and Tool Users
- All that structure matches does not glitter
- Towards Fully Automated Molecular Simulations: Multi-Agent Framework for Simulation Setup and Force Field Extraction
- RoboChemist: Long-Horizon and Safety-Compliant Robotic Chemical Experimentation
- Agents of Discovery
- Language Native Lightly Structured Databases for Large Language Model Driven Composite Materials Research
- The Physical Basis of Prediction: World Model Formation in Neural Organoids via an LLM-Generated Curriculum
- IR-Agent: Expert-Inspired LLM Agents for Structure Elucidation from Infrared Spectra
- Osprey: Production-Ready Agentic AI for Safety-Critical Control Systems
- An AI-driven workflow for the accelerated optimization of cell-free protein synthesis
- AI Agents for Photonic Integrated Circuit Design Automation
- LARC: Towards Human-level Constrained Retrosynthesis Planning through an Agentic Framework
- The Rise of Generative AI for Metal-Organic Framework Design and Synthesis
- Considering the ethics of large machine learning models in the chemical sciences
- What are the limits to biomedical research acceleration through general-purpose AI?
- MiGrATe: Mixed-Policy GRPO for Adaptation at Test-Time
- Multi-agent systems for chemical engineering: A review and perspective
- ToPolyAgent: AI agents for coarse-grained bead-spring topological polymer simulations
- LLM-based Multi-Agent Copilot for Quantum Sensor
- Operationalizing Serendipity: Multi-Agent AI Workflows for Enhanced Materials Characterization with Theory-in-the-Loop
- Expert-Guided LLM Reasoning for Battery Discovery: From AI-Driven Hypothesis to Synthesis and Characterization
- zERExtractor:An Automated Platform for Enzyme-Catalyzed Reaction Data Extraction from Scientific Literature
- SciToolAgent: A Knowledge Graph-Driven Scientific Agent for Multi-Tool Integration
- Autonomous agentic design for photonics
- Reproducibility of Machine Learning-Based Fault Detection and Diagnosis for HVAC Systems in Buildings: An Empirical Study
- Protein design and optimization for synthetic cells
- CyberNeuro: A Privacy-Preserving Agentic Workbench for Cohort-Scale Neuroimage and Clinical Data Analysis
- AI Research Agents for Machine Learning: Search, Exploration, and Generalization in MLE-bench
- Can AI Evaluate AI Scientists? A Benchmarking Study of Autonomous Research Generation Systems Using Automated Multi-Model Review
- BioMARS: A Multi-Agent Robotic System for Autonomous Biological Experiments
- Spatial transcriptomics AI agent charts hPSC-pancreas maturation in vivo
- Exploring Modularity of Agentic Systems for Drug Discovery
- THE-Tree: Can Tracing Historical Evolution Enhance Scientific Verification and Reasoning?
- Active Inference AI Systems for Scientific Discovery
- MolProphecy: Bridging Medicinal Chemists' Knowledge and Molecular Pre-Trained Models via a Multi-Modal Framework
- Large Language Model Agent for Modular Task Execution in Drug Discovery
- CoMind: Towards Community-Driven Agents for Machine Learning Engineering
- A Survey of LLM-Driven AI Agent Communication: Protocols, Security Risks, and Defense Countermeasures
- Federated Learning from Molecules to Processes: A Perspective
- A Large Language Model-based Multi-Agent Framework for Analog Circuits' Sizing Relationships Extraction
- AutoSDT: Scaling Data-Driven Discovery Tasks Toward Open Co-Scientists
- Modeling Earth-Scale Human-Like Societies with One Billion Agents
- AlphaEvolve: A coding agent for scientific and algorithmic discovery
- Perspective on Utilizing Foundation Models for Laboratory Automation in Materials Research
- GeneWhisperer: Enhancing manual genome annotation with large language models
- Intelligent Design 4.0: Paradigm Evolution Toward the Agentic AI Era
- Training a Scientific Reasoning Model for Chemistry
- Soft causal learning for generalized molecule property prediction: An environment perspective
- A VLM-based Method for Visual Anomaly Detection in Robotic Scientific Laboratories
- The rise of agents: Computational chemistry is ready for (R)evolution
- AI Scientists Fail Without Strong Implementation Capability
- ChemAU: Harness the Reasoning of LLMs in Chemical Research with Adaptive Uncertainty Estimation
- EXP-Bench: Can AI Conduct AI Research Experiments?
- Large Language Model-Based Agents for Automated Research Reproducibility: An Exploratory Study in Alzheimer's Disease
- BioCLIP 2: Emergent Properties from Scaling Hierarchical Contrastive Learning
- GAIDeT (Generative AI Delegation Taxonomy): A taxonomy for humans to delegate tasks to generative artificial intelligence in scientific research and publishing
- CPathAgent: An Agent-based Foundation Model for Interpretable High-Resolution Pathology Image Analysis Mimicking Pathologists' Diagnostic Logic
- L-MTP: Leap Multi-Token Prediction Beyond Adjacent Context for Large Language Models
- BioDSA-1K: Benchmarking Data Science Agents for Biomedical Research
- X-MAS: Towards Building Multi-Agent Systems with Heterogeneous LLMs
- Long-Horizon Autonomous Architecture Research with a Language-Model Agent: A Behavioural Case Study
- RN-F: A Novel Approach for Mitigating Contaminated Data in Large Language Models
- ChemPile: A 250GB Diverse and Curated Dataset for Chemical Foundation Models
- AI and biosecurity: The need for governance
- CRAFTS: Collaborative Role-Adaptive Fine-Tuning of LLM Agents for Chemical Process Simulation
- MLLM-based Discovery of Intrinsic Coordinates and Governing Equations from High-Dimensional Data
- XtraGPT: Context-Aware and Controllable Academic Paper Revision via Human-AI Collaboration
- Agentic Bayesian Optimization through Surrogate-Augmented Autoresearch
- Putting It All into Context: Simplifying Agents with LCLMs
- Assessing the Chemical Intelligence of Large Language Models
- LLM-Augmented Chemical Synthesis and Design Decision Programs
- Long-Term Memory for VLA-based Agents in Open-World Task Execution
- Empowering Scientific Workflows with Federated Agents
- Leveraging Large Language Models for enzymatic reaction prediction and characterization
- ChemQuests: A Curated Chemistry Question-Answer Database Extracted from ChemRxiv papers
- MolClaw: An Autonomous Agent with Hierarchical Skills for Drug Molecule Evaluation, Screening, and Optimization
- Inverse Design of Inorganic Compounds with Generative AI
- Vibe Researching as Wolf Coming: Can AI Agents with Skills Replace or Augment Social Scientists?
- An Onto-Relational-Sophic Framework for Governing Synthetic Minds
- Towards a new paradigm of scientific discovery with socialized artificial intelligence
- Deciphering Scientific Reasoning Steps from Outcome Data for Molecule Optimization
- Learning the ARTS of Search for Automated Discovery
- ENPIRE: Agentic Robot Policy Self-Improvement in the Real World
- Cordon: Semantic Transactions for Tool-Using LLM Agents
- InvDesMobility: a reliability-gated first-principles feedback framework for closed-loop materials discovery
- ChargeBD: Character-Aware Heterogeneous Agent Reasoning for Guided Engineering in Battery Development
- Escaping the Hydrolysis Trap: An Agentic Workflow for Inverse Design of Durable Photocatalytic Covalent Organic Frameworks
- AI4S-SDS: A Neuro-Symbolic Solvent Design System via Sparse MCTS and Differentiable Physics Alignment
- A literature-grounded scientific reasoning framework for defect-engineered TiO2 photocatalysts
- MLIPilot: LLM-Driven Auto-Research for Machine-Learned Interatomic Potentials
- Harnessing AtomisticSkills for Agentic Atomistic Research
- DrugSAGE:Self-evolving Agent Experience for Efficient State-of-the-Art Drug Discovery
- APWA: A Distributed Architecture for Parallelizable Agentic Workflows
- Engineering-Oriented Symbolic Regression: LLMs as Physics Agents for Discovery of Simulation-Ready Constitutive Laws
- Computational Reasoning of Large Language Models
- End-to-end autonomous scientific discovery on a real optical platform
- Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning
- Measuring the Gap Between Human and LLM Research Ideas
- PaperClaw: Harnessing Agents for Autonomous Research and Human-in-the-Loop Refinement
- Automated reproducibility assessments in the social and behavioral sciences using large language models
- Can AI Scientist Agents Learn from Lab-in-the-Loop Feedback? Evidence from Iterative Perturbation Discovery
- PhenoAssistant: A Conversational Multi-Agent AI System for Automated Plant Phenotyping
- EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale
- VERITAS: A Multi-Agent Co-Scientist for Verifiable Image-Derived Hypothesis Testing
- Epistemic Blinding: An Inference-Time Protocol for Auditing Prior Contamination in LLM-Assisted Analysis
- Integration of artificial intelligence and big data in materials science: New paradigms and scientific discoveries
- Application of multi-agent in energy materials research
- Autonomous Quantum Simulation through Large Language Model Agents
- Large Language Models to Accelerate Organic Chemistry Synthesis
- CheMLFlow: An Open-Source Platform for Cheminformatics and Materials Informatics Applications
- Revealing the Intrinsic Ethical Vulnerability of Aligned Large Language Models
- SimMOF: AI agent for Automated MOF Simulations
- Language Models for Materials Discovery and Sustainability: Progress, Challenges, and Opportunities
- OPERA: Operator-residual feedback for reliable autonomous optical experiments with language-model agents
- Simulator-Grounded Large Language Models for Industrial Causal Reasoning: Tool-Use, Structured Injection, and Plant-Portable Retrieval for Wastewater Treatment Decision Support
- MooseAgent: A LLM Based Multi-agent Framework for Automating Moose Simulation
- Benchmarking large language models for materials synthesis: the case of atomic layer deposition
- Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality
Related