# Reinforcement Learning GUI Positioning

GTA1 is a state-of-the-art GUI positioning model trained based on Reinforcement Learning (GRPO). It achieves precise interface element positioning by directly motivating actionable responses rather than lengthy inferences.

GTA1 is a GUI positioning model based on Reinforcement Learning (GRPO). It achieves precise positioning by directly rewarding successful clicks, avoiding lengthy thought-chain reasoning.