Jev-Mobile speeds up mobile app control by splitting planning and action
Jev-Mobile: Jev as an Executor for Mobile GUI Agents
Artificial IntelligenceSoftware Engineering
Summary
Using AI to control mobile apps often means a large model plans and acts at every step, which can be slow and costly. The researchers made Jev-Mobile, which only uses the big AI for planning some of the time and uses a faster method to handle simple app actions in between. This way, it completes tasks nearly as well but much faster and cheaper. The system was tested on many Android tasks and showed strong results.
What this means in practice
- •For mobile app developers: Build AI assistants that control mobile apps more quickly and cost-effectively by separating high-level planning from low-level action steps.
- •For software engineers: Develop responsive GUI automation tools that reduce reliance on costly large models for every action by using lightweight decision mechanisms.
Authors
Linghua Zhang
Abstract
Vision-language models (VLMs) have become a common foundation for autonomous mobile GUI agents, but most existing systems rely on the VLM for both planning and action grounding at nearly every interaction step, leading to substantial latency and model-serving cost. We introduce Jev-Mobile, which shifts this paradigm to low-frequency VLM planning and high-frequency lightweight execution: the VLM specifies local goals, the accessibility tree defines a structured executable action space, and Jev, a fast typed decision model, repeatedly selects actions within this space. This design allows multiple GUI actions to be executed under a single VLM decision, reducing expensive VLM inference while preserving adaptive interaction. On the full AndroidWorld task suite, Jev-Mobile achieves 79% task success, compared with 78% for SeeAct-V and 84% for a Step-wise VLM baseline. Among successful trajectories, it reduces mean end-to-end execution time by 32.7% and mean model API cost by 73.4% relative to Step-wise VLM. These results show that decoupling high-level VLM reasoning from low-level action execution can substantially improve mobile GUI agent efficiency while maintaining competitive task performance.