@xywang626
We are super excited to release OpenCUA โ the first from 0 to 1 computer-use agent foundation model framework and open-source SOTA model OpenCUA-32B, matching top proprietary models on OSWorld-Verified, with full infrastructure and data. ๐ [Paper] https://t.co/SYEio5ccNJ ๐ [Website] https://t.co/ma6bBuYiNM ๐ค [Models] https://t.co/7TVtIdjkmq ๐[Data] https://t.co/N6tQQwQkhs ๐ป [Code] https://t.co/ihr8TXmG6k ๐ OpenCUA โ comprehensive open-source framework for computer-use agents, including: ๐ AgentNet โ first large-scale CUA dataset (3 systems, 200+ apps & sites, 22.6K trajectories) ๐ OpenCUA model โ open-source SOTA on OSWorld-Verified (34.8% avg success, outperforms OpenAI CUA) ๐ฅ AgentNetTool โ cross-system computer-use task annotation tool ๐ AgentNetBench โ offline CUA benchmark for fast, reproducible evaluation ๐ก Why OpenCUA? Proprietary CUAs like Claude or OpenAI CUA are impressive๐คฏ โ but thereโs no large-scale open desktop agent dataset or transparent pipeline. OpenCUA changes that by offering the full open-source stack ๐ : scalable cross-system data collection, effective data formulation, model training strategy, and reproducible evaluation โ powering top open-source models including OpenCUA-7B and OpenCUA-32B that excel in GUI planning & grounding. Details of OpenCUA framework๐