MobileGUI-VBench Dataset Documentation

A Benchmark for Mobile GUI Agent Evaluation · Statistics & Visualization

150
Total Tasks
40
Unique Apps
8
Scenario Categories

1. Overview

MobileGUI-VBench is a benchmark suite for evaluating mobile GUI Agent capabilities. It covers a wide range of real-world scenarios including Social & Communication, Music & Radio, Video Streaming, Shopping & Deals, Travel & Transit, Navigation, Lifestyle Services, and News & Reading. The benchmark encompasses single-app and cross-app tasks, single-intent and multi-intent tasks, explicit instructions and implicit-intent instructions, and is designed to evaluate an Agent's instruction comprehension and execution capability in real mobile app environments.

Data files: MobileGUI-VBench.xlsx (full annotations), MobileGUI-VBench.jsonl (reduced field set)

2. Column Descriptions

ColumnDescription
TASK_IDUnique task identifier
InstructionRaw user query / instruction text
Scenario CategoryTask scenario category (e.g., Social & Communication, Music & Radio, Video Streaming)
APPApp(s) involved in the task, multiple apps separated by 「、」
Instruction TypeAgent-Execution / Agent-User Interaction
Chain ComplexitySimple / Medium / Complex
Operation TypeBasic Operation / Advanced Operation
Instruction ClarityExplicit Instruction / Implicit Intent Instruction / Ambiguous Instruction
Intent CompositionSingle Intent / Dependent Multi-Intent / Independent Multi-Intent
Function PointSpecific functionality being tested (e.g., Search, Product Lookup, Watch Video)
PreconditionRequired state or environment setup for task completion
Environment PresetSpecific environment setup steps (provided for some tasks)
Ideal ClarificationExpected clarification behavior from the Agent (mostly for Agent-User Interaction tasks)

3. Tag Definitions

This section defines the multi-dimensional tags used in MobileGUI-VBench, serving as the criteria for task annotation and result analysis. The tag system consists of 6 classification dimensions (Instruction Type / Chain Complexity / App Scope / Operation Type / Instruction Clarity / Intent Composition) and 2 auxiliary fields (Precondition / Ideal Clarification).

3.1 Instruction Type

TagDefinitionExample
Agent-ExecutionThe instruction is complete and unambiguous; the Agent can execute it directly without asking the user"Set my QQ status to Busy"
Agent-User InteractionThe instruction lacks sufficient information or has multiple valid execution paths; the Agent is expected to proactively clarify with the user before executing"Change my QQ status to something else" (needs clarification: to which status?)

3.2 Chain Complexity

Categorized by the estimated number of steps required to complete the task (each tap / input / swipe counts as one step):

TagDefinitionExample
SimpleCompletable within ≤ 5 steps"Open NetEase Cloud Music"
MediumRequires 6 ~ 10 steps"Search and play a tech video on Bilibili"
ComplexRequires > 10 steps, or involves deep multi-page navigation"Find the top-rated coffee shop on RED and save the post"

3.3 App Scope

TagDefinitionExample
Single-AppThe task is completed entirely within one app"Play my favorite songs in QQ Music"
Cross-AppThe task requires switching between ≥ 2 apps or transferring data across apps"Open in Amap the location my friend sent me in WeChat"

3.4 Operation Type

TagDefinitionUI Actions Involved
Basic OperationUses only the most common GUI interactionsTap, text input
Advanced OperationInvolves more complex gestures or control manipulationSwipe, drag, long-press, double-tap, pinch-zoom, scroll-targeting

3.5 Instruction Clarity

TagDefinitionExample
Explicit InstructionThe instruction clearly specifies the target, action, and parameters; no inference required"Pin my chat with Zhang San in WeChat"
Implicit Intent InstructionThe instruction does not explicitly state the real intent; the Agent must infer it from common sense or context"Haven't been on QQ for a while, see what my friends have been posting" (implicit intent: view friend feeds)
Ambiguous InstructionThe instruction lacks key parameters or has multiple valid interpretations; the Agent must clarify"Help me search for that movie" (which movie? on which platform?)

3.6 Intent Composition

TagDefinitionExample
Single IntentThe instruction contains only one independent task"Open NetEase Cloud Music and play my favorite playlist"
Dependent Multi-IntentThe instruction contains multiple tasks with conditional dependency or data passing between them (B requires the result of A)"Search for hotpot restaurants near me on Amap, then send the nearest one to Zhang San on WeChat" (B depends on A's output)
Independent Multi-IntentThe instruction contains multiple tasks that are independent and can be executed in parallel"Set my QQ status to Busy, and also open NetEase Cloud Music"

3.7 Precondition

The environment state or account state required before executing the task. An empty field means no special precondition is needed.

Examples:

3.8 Ideal Clarification

For Agent-User Interaction tasks, this field defines the clarification or confirmation behavior the Agent is expected to proactively initiate. It is the core basis for evaluating the Agent's proactive interaction capability.

Examples:

4. Scenario Category Distribution

Scenario CategoryCountPercentage
Video Streaming2416.0%
Shopping & Deals2416.0%
Lifestyle Services2315.3%
Social & Communication2114.0%
Travel & Transit1812.0%
Music & Radio1711.3%
Navigation1711.3%
News & Reading64.0%

5. App Distribution

The dataset involves 40 unique apps. A single task may involve multiple apps, so the total count of app occurrences exceeds the number of tasks (150).

All 40 apps shown, sorted by occurrence frequency (descending).

6. Instruction Type Distribution

Instruction TypeCountPercentage
Agent-Execution9160.7%
Agent-User Interaction5939.3%

7. Chain Complexity Distribution

Chain ComplexityCountPercentage
Medium9060.0%
Simple3624.0%
Complex2416.0%

8. Operation Type Distribution

Operation TypeCountPercentage
Basic Operation12382.0%
Advanced Operation2718.0%

9. Instruction Clarity Distribution

Instruction ClarityCountPercentage
Explicit Instruction7650.7%
Implicit Intent Instruction6744.7%
Ambiguous Instruction74.7%

10. Intent Composition Distribution

Intent CompositionCountPercentage
Single Intent9362.0%
Dependent Multi-Intent4932.7%
Independent Multi-Intent85.3%

11. Function Point Distribution (Top 15 of 77)

The function points are diverse (77 types in total). Only the top 15 by frequency are shown here. The remaining function points mostly appear 1–2 times, covering more granular long-tail operation scenarios.

12. Summary

The MobileGUI-VBench dataset contains 150 tasks across 40 mainstream mobile apps and 8 real-world scenario categories. The task design balances single-app and cross-app collaboration, single-intent and multi-intent workflows, and explicit, implicit, and ambiguous instruction comprehension, making it a comprehensive benchmark for evaluating the scenario generalization and complex instruction execution capabilities of mobile GUI Agents.