A lightweight, powerful framework for multi-agent workflows

Muon is an optimizer for hidden layers in neural networks

Benchmarking the Spectrum of Agent Capabilities

"AutoAgent: Fully-Automated and Zero-Code LLM Agent Framework"

A collection of model counting (#SAT) benchmarks.

[NeurIPS 2023] Tree of Thoughts: Deliberate Problem Solving with Large Language Models

Fully open reproduction of DeepSeek-R1

PyTorch native post-training library

Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

✨✨Latest Advances on Neuro-Symbolic Learning in the era of Large Language Models

SatLM: SATisfiability-Aided Language Models using Declarative Prompting (NeurIPS 2023)

A toolkit for SAT-based prototyping in Python

The glucose SAT solver

A minimalistic and high-performance SAT solver

Get started with building Fullstack Agents using Gemini 2.5 and LangGraph

Prover9 is an automated theorem prover for first-order and equational logic, and Mace4 searches for finite models and counterexamples.

Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek-R1, Qwen3, Gemma 3, TTS 2x faster with 70% less VRAM.

Implementing DeepSeek R1's GRPO algorithm from scratch

A very simple GRPO implement for reproducing r1-like LLM thinking.

[NeurIPS 2025] Thinkless: LLM Learns When to Think

LogicBench is a natural language question-answering dataset consisting of 25 different reasoning patterns spanning over propositional, first-order, and non-monotonic logics.

(ACL 2025 Main) Code for MultiAgentBench : Evaluating the Collaboration and Competition of LLM agents https://www.arxiv.org/pdf/2503.01935

Framework and Language for Neurosymbolic Programming.

One repository is all that is necessary for Multi-agent Reinforcement Learning (MARL)

An open collection of methodologies to help with successful training of large language models.