OpenPipe/ART

open-source RL framework for training LLM agents on real-world tasks

Build and train custom RL environments for LLM agents with modular tools and easy integration
willccbb/verifiers
Open cookbook for building top-tier code LLMs with complete training data and reproducible pipeline
OpenCoder-llm/OpenCoder-llm
open format for guiding AI coding agents with project instructions
openai/agents.md
Microsoft’s open-source foundation model for multimodal AI agents
microsoft/Magma
Task engine for AI agents
lmnr-ai/flow
open-source reinforcement learning for language models
agentica-project/rllm
multi-agent LLM app framework with modular tools and real-time control
agentscope-ai/agentscope
framework for building LLM and RAG apps with pipelines
deepset-ai/haystack
open-source coding assistant with agentic features, supports Claude, GPT-4o, Gemini, and local LLMs
amrit110/oli
framework for building and running autonomous agents with open-source models
Tencent/Youtu-agent