Current benchmarks for graphical user interface (GUI) agents predominantly rely on static screenshots. However, real-world smartphone interaction routinely requires agents to process transient audio cues and temporal video dynamics that are tightly coupled with the moment of action. To bridge this gap, we introduce OmniGUI, the first step-level benchmark designed to evaluate GUI agents in omni-modal smartphone environments. OmniGUI provides continuous, interleaved multimodal inputs-comprising static images, synchronous audio, and video clips—at every action step. The dataset encompasses 708 expert-demonstrated episodes (2,572 action steps) across 29 applications, systematically annotated with objective multimodal dependency levels. Because dedicated omni-modal GUI agent frameworks are currently in their nascent stage, we select foundational omni-modal models capable of natively processing interleaved inputs to serve as agent proxies for our initial baselines. Our empirical evaluation reveals that while current models exhibit competency on visually static tasks, their action prediction performance degrades significantly in environments requiring synchronous temporal and auditory signals. Furthermore, ablation studies isolate specific operational bottlenecks, notably cross-modal interference when processing task-irrelevant environmental noise.

OmniGUI Framework Overview: A GUI agent completes smartphone tasks across multiple steps based on multimodal instructions. At each step, it processes synchronous inputs—static screenshots, real-time audio, temporal video clips, and action history—to predict the next action (e.g., TYPE, TAP). Performance is evaluated using Type Match (TM) and Exact Match (EM) against human demonstrations.