Forget Simple Prompts: This AI's Proactive Context Management Unlocks Long-Horizon Mobile Tasks

Current MLLM-based mobile GUI agents struggle with long-horizon tasks that require retaining intermediate facts across many steps and app transitions. The root cause is ReAct-style prompting, which passively accumulates per-step records, leading to prompt explosion and dilution of critical cross-app facts. To solve this, researchers from multiple institutions introduce MemGUI-Agent, built on a novel paradigm called Context-as-Action (ConAct). ConAct treats context management as first-class actions emitted by the same policy that selects UI actions, instead of passively appending history. It maintains three structured context fields: folded action history, folded UI state, and recent step record, preserving critical UI facts while keeping context compact.

To make proactive context management learnable across model scales, the team constructed MemGUI-3K, a dataset of 2,956 trajectories with full ConAct annotations for supervised training and offline analysis. Training an 8B model on this dataset produced MemGUI-8B-SFT, which achieves the best open-data 8B performance on MemGUI-Bench and generalizes to the out-of-distribution MobileWorld benchmark. The approach demonstrates that proactive context management significantly outperforms passive history accumulation for long-horizon mobile tasks, enabling reliable automation across multiple app switches without losing critical context. Code, data, and trained models will be released publicly.