Search papers, labs, and topics across Lattice.
MobileGym is a new browser-based simulation environment for mobile GUI agents that provides verifiable outcome signals and scalable online RL through parallel rollouts. The environment uses structured JSON to represent and manipulate the full environment state, enabling deterministic state-based judging and dense RL rewards. Experiments show that agents trained in MobileGym can transfer to real-world devices, with GRPO on Qwen3-VL-4B-Instruct gaining +12.8 percentage points on a 256-task test set and retaining 95.1% of the simulation-side training gain on a 59-task real-device subset.
Training agents in MobileGym transfers surprisingly well to real-world mobile devices, retaining over 95% of the simulation-side performance gains.
We present MobileGym, a browser-hosted, lightweight, fully controllable environment for everyday mobile use, targeting interaction fidelity without replicating proprietary backends. It enables two capabilities previously out of reach for everyday apps: verifiable outcome signals through deterministic state-based judging over structured JSON state, and scalable online RL through low-cost parallel rollouts. The full environment state is captured, configured, forked, and compared as structured JSON, and a single server can host hundreds of parallel instances, with about 400 MB memory per instance and about 3 s cold start. A layered state model and a declarative task-definition framework keep state programmability and task creation practical at scale, and a single programmatic judging mechanism delivers both deterministic evaluation verdicts and dense RL rewards. The accompanying MobileGym-Bench provides 416 parameterized task templates, including 256 test and 160 train templates, over 28 apps, with deterministic judges and a structured AnswerSheet protocol that avoids free-text matching failures. In a Sim-to-Real case study, GRPO on Qwen3-VL-4B-Instruct gains +12.8 percentage points on the 256-task test set, and on a 59-task real-device signal subset, real-device execution retains 95.1% of the simulation-side training gain. Project page: https://mobilegym.github.io.
