Stars
CIFAR-10 speedruns: 94% in 2.6 seconds and 96% in 27 seconds
A PyTorch library and evaluation platform for end-to-end compression research
Post-training with Tinker
A Co-evolving Agentic AI System for Medical Imaging Analysis
[NeurIPS 2025] TTRL: Test-Time Reinforcement Learning
A fast and flexible Python package for efficiently solving lasso, elastic net, group lasso, and group elastic net problems.
Open-source implementation of AlphaEvolve
KernelBench: Can LLMs Write GPU Kernels? - Benchmark with Torch -> CUDA problems
Formalization of the Millennium Problems in Lean4.
gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI
Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen2.5, Qwen3, Llama, and more!
Kimina Lean server (+ client SDK)
Repo for Rho-1: Token-level Data Selection & Selective Pretraining of LLMs.
Aioli: A unified optimization framework for language model data mixing
Official Repository of Absolute Zero Reasoner
Environments for LLM Reinforcement Learning
[NeurIPS 2025 Spotlight] Reasoning Environments for Reinforcement Learning with Verifiable Rewards
[NeurIPS 2025] Reinforcement Learning for Reasoning in Large Language Models with One Training Example
[NeurIPS 2025] What Makes a Reward Model a Good Teacher? An Optimization Perspective
This repo contains the dataset and code for the paper "SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering?"
Ongoing research training transformer models at scale