Odyssey: Empowering Minecraft Agents with Open-World Skills
✨Odyssey is a framework that empowers LLM-based Minecraft agents with open-world skills, featuring 40 primitive skills and 183 compositional skills, enabling AI to autonomously explore, learn, and execute diverse tasks in the Minecraft universe.
Agent & ToolingPythonLangChain
ART (Agent Reinforcement Trainer)
🧠An open-source reinforcement learning framework that trains multi-step agents for real-world tasks using GRPO, supporting Qwen2.5, Qwen3, Llama, and other large language models.
Model & Inference FrameworkPythonvLLM
BigCodeBench
🧠A benchmark for evaluating the code generation capabilities of large language models, featuring 1,140 software-engineering-oriented programming tasks with two modes (Complete and Instruct) to test models on complex instructions and diverse function call scenarios.
Docs, Tutorials & ResourcesPythonPyTorch
StateSpace
✨A modern framework for probabilistic programming and Bayesian analysis designed for research and data analytics, featuring intuitive APIs and flexible model definition capabilities。
Model & Inference FrameworkPythonMachine Learning
AgentGym
✨A comprehensive platform for training, evaluating, and evolving LLM-based agents across diverse environments with standardized benchmarks.
Agent & ToolingPythonPyTorch
mario-ai
✨A reinforcement learning environment for Mario AI, offering trainable agents to play Super Mario games.
Agent & ToolingPythonPyTorch
Mini Agent
✨A minimal yet professional demo project that showcases the best practices for building agents with the MiniMax M2.1 model, featuring a complete execution pipeline and production-grade capabilities.
Agent & ToolingPythonAI Agents
how-to-build-a-coding-agent
✨A step-by-step workshop that teaches you how to build your own AI-powered coding assistant, starting from a basic chatbot and progressively adding powerful tools like file reading, shell command execution, and code search.
Docs, Tutorials & ResourcesGoAI Agents
MiroThinker
✨An open-source deep research agent optimized for research and prediction tasks, achieving 80.8% Avg@8 score on the challenging GAIA benchmark, featuring 256K context window support and up to 600 tool calls per task.
Agent & ToolingPythonAI Agents
learn-claude-code
✨An educational project that teaches you how to build modern AI coding agents from scratch through progressive tutorials, featuring 5 versions from simple bash tools to a complete skills system.
Docs, Tutorials & ResourcesPythonAI Agents