2026-08-18
We are happy to announce Lego-RL: an open-source framework for harness-native reinforcement learning of coding agents. All code, data, and models are available on GitHub and Hugging Face.
LegoX
An open-source collection advancing frontier agentic intelligence
LegoX is a series of open-source projects spanning agentic data pipelines, training frameworks, evaluation benchmarks and foundation models. Our mission is to connect these pieces into reproducible and extensible systems that advance frontier intelligence.
News
Hugging Face2026-08-01
Congratulations! SWE-Lego models and datasets have surpassed 200K cumulative downloads on Hugging Face.
SWE-Review
An agentic code-review system that diagnoses repository-level problems and closes the generate–review–revise loop.
Read project →
Terminal-Lego
A study of interaction trajectories and environment-grounded supervision for building more capable terminal agents.
Read project →
SWE-Lego
A supervised fine-tuning recipe and model family for resolving real software issues with carefully validated trajectories.
Read project →
Collaboration
Collaborating institutions
