StepReflect: Structured UI Transition Reflection for Mobile GUI Agents
For developers of mobile GUI agents, StepReflect offers a cost-effective, locally deployable alternative to expensive frontier-model reflection, improving reliability in long-horizon tasks.
StepReflect introduces a structured prediction approach for per-step GUI state transition reflection in mobile agents, achieving 82.16% transition-level accuracy on AndroidWorld (11.83 points above zero-shot GPT-5.2) and improving task success in three of four agent configurations while reducing API costs.
Autonomous mobile GUI agents require accurate action reflection for reliable long-horizon execution. Existing approaches rely on open-ended multimodal reasoning after each action, which is costly and poorly matched to the structured nature of GUI state transitions. We propose StepReflect, which formulates per-step GUI reflection as supervised structured prediction conditioned on explicit transition specifications and paired visual evidence. StepReflect is trained through a staged pipeline combining supervised fine-tuning, teacher-student distillation, and preference- and reward-based refinement. Offline, the resulting 8B model achieves 82.16% transition-level accuracy on AndroidWorld, exceeding zero-shot GPT-5.2 by 11.83 percentage points under the same structured input. Online, across M3A, Agent-SAMA, MAI-UI-8B, and Seed-2.0-Pro, StepReflect achieves higher task success in three of four agent configurations and remains within one successful task of the GPT-5.2 Reflection Agent in the fourth. It also reduces paid API charges relative to GPT-based reflection in all four configurations. These results establish StepReflect as a practical, locally deployable alternative to repeated frontier-model reflection for long-horizon mobile GUI agents.