🐧 Harness for RSI. AI Build Better AI

Agent World Model: Infinity Synthetic Environments for Agentic Reinforcement Learning

Wrap Antigravity, ChatGPT Codex, Claude Code, Grok Build as an OpenAI/Gemini/Claude/Codex compatible API service, allowing you to enjoy the free Gemini 3.1 Pro, GPT 5.6 Series, Grok 4.5, Claude mod…

Unofficial implementation of the Dreamer 4 world model in PyTorch.

Lets make video diffusion practical!

Environment-Native Verified Search (ENVS): GUI agent post training pipeline that reaches higher accuracy at lower compute than online RL.

Research and development (R&D) is crucial for the enhancement of industrial productivity, especially in the AI era, where the core aspects of R&D are mainly focused on data and models. We are commi…

[ICML 2026] What Does Vision Tool-Use Reinforcement Learning Really Learn? Disentangling Tool-Induced and Intrinsic Effects for Crop-and-Zoom

Code for paper OpenWebRL: Online Multi-Turn Reinforcement Learning for Visual Web Agents

InfoSFT is a modified supervised fine-tuning algorithm that generalizes better and forgets less.

CUA-Gym-Hub: mock web apps as reproducible RL training environments for computer-use agents

Scalable pipeline for synthesizing verifiable RLVR training data for computer-use agents

Symphony turns project work into isolated, autonomous implementation runs, allowing teams to manage work instead of supervising coding agents.

OpenSeeker: A search agent with open-source data and models

[NeurIPS 2025 Spotlight] OpenCUA: Open Foundations for Computer-Use Agents

Code for "WebVoyager: WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models"

An Illusion of Progress? Assessing the Current State of Web Agents

Windows Agent Arena (WAA) 🪟 is a scalable OS platform for testing and benchmarking of multi-modal AI agents.

This is the official code base of AgentNetTool in OpenCUA. Website: https://opencua.xlang.ai/

SWE-bench: Can Language Models Resolve Real-world Github Issues?

[NeurIPS 2024] OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments

Implementation of paper: Scaling the Scaling Logic

The official repo for “Dolphin: Document Image Parsing via Heterogeneous Anchor Prompting”, ACL, 2025.

[ICLR 2026] Official code for TraceRL: Revolutionizing post-training for Diffusion LLMs, powering the SOTA TraDo series.

Easy and Efficient dLLM Fine-Tuning

DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation

MMaDA - Open-Sourced Multimodal Large Diffusion Language Models (dLLMs with block diffusion, mixed-CoT, unified RL)