MDL-7765 EST.2024 · IDX.528 Science In production ⇄ Compare
AlphaFold 3 Google DeepMind · USA / UK · 2024
The Nobel-recognized model predicting the structure of life's molecules and their interactions.
AlphaFold 3 predicts the joint structure of proteins, DNA, RNA, ligands and antibodies — the interactions that underlie disease and drug action. Following AlphaFold 2, whose creators received the 2024 Nobel Prize in Chemistry, it is used by millions of researchers and anchors a new era of AI-driven drug discovery through Isomorphic Labs.
#biology #Nobel Prize #drug discovery #protein structure
Official website ↗ ▸ SpecificationsClass & identity
Model class Scientific structure-prediction model
Version / release 2024 Architecture
Architecture Diffusion-based (pairformer/generative)
Parameters —
Context window —
Modalities Sequences → 3D molecular structures
Knowledge cutoff / data — Performance (benchmarks)
MMLU —
GPQA Diamond —
SWE-bench —
MATH / AIME —
Chatbot Arena Elo —
Domain benchmark SOTA joint structure prediction Capabilities Vision
Reasoning mode —
Tool use / agentic —
Multilinguality —
Key strengths Proteins, DNA, RNA, ligands, antibodies Access & deployment
Availability Server (non-commercial) + Isomorphic
License Restricted (non-commercial server)
Pricing Not sold. AlphaFold Server is free of charge for non-commercial users, and the weights are released only on request, for non-commercial research.
Deployment Web server / partner
Safety / alignment Access controls Compiled to the wujec.ai Spec Standard v1 — “—” means not publicly disclosed.