#2121st of 42 in Open modelsFor science, health and research
alphafold3
Inference pipeline for AlphaFold 3 biomolecular structure prediction
- Stars
- 8.7k
- License
- Apache-2.0
- Last commit
- Oct 2026
- Last release
- Jul 2026
- Language
- Python
Missed the latest GitHub check; these facts may be older.
Overview
Runs the AlphaFold 3 inference pipeline, predicting structures of proteins, nucleic acids, ligands and their complexes from a JSON input. The pipeline has a CPU-only genetic and template search stage and a GPU inference stage, which can be run separately. Model parameters are not included and must be requested from Google under separate terms of use.
Who it is for: Structural biology researchers running AlphaFold 3 predictions on their own GPU
Strengths
- Code is Apache-2.0 and covers the full inference path from JSON input to structure output
- Data pipeline and inference can run separately via --run_data_pipeline and --run_inference
- Alternative weights offered: SynthID Bio watermarked and Leaving Atom fine-tuned variants
- Built on JAX and Haiku, with separate input, output and performance docs
Weaknesses
- Weights are not open: use only if received directly from Google, under terms of use
- Requires separate genetic databases to be downloaded; sizes not stated in the README
- Inference needs a GPU; the README gives no VRAM or RAM figures
- Only the inference pipeline is included, no training code
What it needs
- GPU required
- Needs Docker, Model parameters from Google, Genetic sequence databases (BFD, MGnify, UniRef90, UniProt, PDB, NT, RFam, RNACentral), HMMER
- Models: AlphaFold 3, AlphaFold 3 SynthID Bio, AlphaFold 3 Leaving Atom
Also in Open models
See all 42| Rank | Project | Score |
|---|---|---|
| 1 | llmfitRust tool that checks which LLMs fit your RAM, CPU and GPU | 87 out of 100 |
| 2 | layaSingle-pass multilingual classifier for typed choice, score and yes/no decisions | 86 out of 100 |
| 3 | timesfmPretrained time-series foundation model for zero-shot forecasting from Google Research | 76 out of 100 |
| #4 | speechPyTorch toolkit for training and deploying ASR, TTS and speech LLM models | 73 out of 100 |
| #5 | yueLyrics-to-song model that writes an editable melody and chord score first | 69 out of 100 |