Fred Wu

OSWorld-V2. OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks

FineVLA. Scalable annotation pipeline for action-aglined fine-grained instruciton for Visual-language-Action model

TestTimeTrainingPapers.

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

OSWorld-G. [NeurIPS 2025 Spotlight] Scaling Computer-Use Grounding via UI Decomposition and Synthesis

AgentTrek. [ICLR2025 Spotlight] Agent Trajectory Synthesis via Guiding Replay with Web Tutorials

yt_chapters. Download YT video with chapters

xlang-paper-reading. Paper collection on building and evaluating language model agents via executable language grounding

Spider2-V. [NeurIPS 2024] Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?

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

threes-ai. AI for the game Threes!