New research introduces UI-TARS, an end-to-end GUI agent model that outperforms existing frameworks by achieving state-of-the-art results in over 10 benchmarks through enhanced perception, unified action modeling, system-2 reasoning, and iterative training with reflective online… https://t.co/MFiQsPTO3w
GUI agent for automated interactions https://t.co/0lT6axhsgn
🏷️:UI-TARS: Pioneering Automated GUI Interaction with Native Agents 🔗:https://t.co/smdiAonVbA https://t.co/VdNBvZzpxG
ByteDance has released UI-TARS, a series of native GUI agent models designed to enhance automated interactions. The models integrate perception, reasoning, grounding, and memory into a unified system. Available in three sizes—2B, 7B, and 72B parameters—UI-TARS aims to outperform existing frameworks, achieving state-of-the-art results across more than 10 benchmarks through improved perception and iterative training techniques. The model is licensed under Apache 2.0, making it accessible for further development and research.