google-deepmind

This repository contains implementations and illustrative code to accompany DeepMind publications

Multi-Joint dynamics with Contact. A general purpose physics simulator.

Open source code for AlphaFold 2.

TensorFlow-based neural network library

AlphaFold 3 inference pipeline.

StarCraft II Learning Environment

Gemma open-weight LLM library, from Google DeepMind

OpenSpiel is a collection of environments and algorithms for research in general reinforcement learning and search/planning in games.

Google DeepMind's software stack for physics-based simulation and Reinforcement Learning environments, using MuJoCo.

A library of reinforcement learning components and agents

A collection of high-quality models for the MuJoCo physics engine, curated by Google DeepMind.

JAX-based neural network library

GDM Science Skills to speed up agentic scientific workflows with better grounding and higher token efficiency. Integrate insights from AlphaGenome, AFDB, UniProt and 30+ other databases and tools.

Monte Carlo tree search in JAX

Optax is a gradient processing and optimization library for JAX.

An open-source library for GPU-accelerated robot learning and sim-to-real transfer.

This API provides programmatic access to the AlphaGenome model developed by Google DeepMind.

Tracking Any Point (TAP)

This dataset code generates mathematical question and answer pairs, from a range of question types at roughly school-level difficulty.

Real-time behaviour synthesis with MuJoCo, using Predictive Control

A library for generative social simulation

bsuite is a collection of carefully-designed experiments that investigate core capabilities of a reinforcement learning (RL) agent

GPU-optimized version of the MuJoCo physics simulator, designed for NVIDIA hardware.

RL research on Android devices.

A collection of formalized statements of conjectures in Lean.

Dramatron uses large language models to generate coherent scripts and screenplays.

tree is a library for working with nested data structures

A platform for managing machine learning experiments

A suite of test scenarios for multi-agent reinforcement learning.

An implementation of the Fermionic Neural Network for ab-initio electronic structure calculations

Research code accompanying AlphaGenome

Reverb is an efficient and easy-to-use data storage and transport system designed for machine learning research

official code for "Large Language Models as Optimizers"

Accompanying code for "Discovering State-of-the-art Reinforcement Algorithms" Nature publication

TORAX: Tokamak transport simulation in JAX

Benchmarking long-form factuality in large language models. Original code for our paper "Long-form factuality in large language models".

Open weights language model from Google DeepMind, based on Griffin.

On the Theoretical Limitations of Embedding-Based Retrieval

Grandmaster-Level Chess Without Search

TIPSv2 (CVPR'26) and TIPS (ICLR'25)

Library to compute surface distance based performance metrics for segmentation tasks.

Minimal and scalable research codebase in JAX, designed for rapid iteration on frontier research in LLM and other autoregressive models.

DQN Zoo is a collection of reference implementations of reinforcement learning agents developed at DeepMind based on the Deep Q-Network (DQN) agent.