Original Apollo 11 Guidance Computer (AGC) source code for the command and lunar modules.

Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents

Official code base for LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels

AI agents running research on single-GPU nanochat training automatically

MOSAIC: A Unified Platform for Cross-Paradigm Comparison and Evaluation of Homogeneous and Heterogeneous Multi-Agent RL, LLM, VLM, and Human Decision-Makers

GLM-Image: Auto-regressive for Dense-knowledge and High-fidelity Image Generation.

Minimal Claude Code alternative. Single Python file, zero dependencies, ~250 lines.

Quick illustration of how one can easily read books together with LLMs. It's great and I highly recommend it.

Native Multimodal Models are World Learners

A tiny CPU simulator written in Python

The best ChatGPT that $100 can buy.

Code and implementations for the ACL 2025 paper "AgentGym: Evolving Large Language Model-based Agents across Diverse Environments" by Zhiheng Xi et al.

Code and implementations for the paper "AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning" by Zhiheng Xi et al.

Making a mini version of the BDX droid. https://discord.gg/UtJZsgfQGe

Our library for RL environments + evals

Chain-of-Agents: End-to-End Agent Foundation Models via Multi-Agent Distillation and Agentic RL.

Render any git repo into a single static HTML page for humans or LLMs

Reference PyTorch implementation and models for DINOv3

gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI

Wan: Open and Advanced Large-Scale Video Generative Models

Hierarchical Reasoning Model Official Release

bluetooth mesh chat, IRC vibes

AIRA-dojo: a framework for developing and evaluating AI research agents

Atropos is a Language Model Reinforcement Learning Environments framework for collecting and evaluating LLM trajectories through diverse environments

An agent benchmark with tasks in a simulated software company.

The Automated LLM Speedrunning Benchmark measures how well LLM agents can reproduce previous innovations and discover new ones in language modeling.

We perform functional grounding of LLMs' knowledge in BabyAI-Text