MobileWorld (Tongyi-MAI)
A containerized Android evaluation environment with GUI-only, agent-user interaction and MCP-augmented tasks, plus checks on application and backend state.
AI-assisted editorial review against a primary source. This is a catalog annotation, not an independent reproduction or endorsement.
What is this resource?
- Resource type
- Benchmarks
- Environment
- android
- Topics
- evaluation, mobile-navigation, human-in-the-loop, tool-integration
- Record publication date
- Not recorded
- Evidence retrieved
- 2026-10-17T09:12:16Z
- Resource ID
- github:tongyi-mai/mobileworld
What licenses are recorded?
Code, training data, weights, and documents can have different terms. A blank declaration means unknown, not unrestricted use.
- Code
- Apache-2.0 (Repository code license; self-hosted applications, model APIs and MCP services keep their own terms.)
- Data
- Not recorded
- Document
- Not recorded
- Weights
- Not recorded
License review notes
- Code: Repository code license; self-hosted applications, model APIs and MCP services keep their own terms. Source 1
Evaluation protocol
- Focus
- Android task completion, agent-user interaction, hybrid GUI and MCP use
- Setting
- interactive
- Observations
- Android application interface, user-agent interaction when enabled
- Actions / outputs
- Android GUI operations, MCP calls when enabled, agent-user interaction when enabled
- Evaluation
- The publisher documents text-answer matching, backend database verification, local-storage inspection and application callbacks.
Scale and counting units
- 201 tasks — README benchmark overview source
- 20 mobile applications — README benchmark overview source
Publisher-reported inventory. GUI-only, agent-user and MCP-augmented configurations have different dependencies and task eligibility; preserve enabled modes and step budget with results.
Compare source-reviewed profiles
Recorded specifications
- availability
- Public task definitions, evaluator and container orchestration source.
- leaderboard url
- https://tongyi-mai.github.io/MobileWorld/#leaderboard
- repository revision
- e7eee8981135ecbe39005a231edf63574defd79c
Explore related topics
Scope and limitations
This is Tongyi-MAI/MobileWorld, distinct from jacklishufan/MobileWorld (MobileWorldBench). Public source includes task definitions, verification logic and environment orchestration. The README reports 201 tasks across 20 applications; this is not an independent count of the evolving task tree. Linux/KVM or WSL2/KVM is required by the documented host setup, with macOS support in progress. MCP and simulated-user settings can require separate external services.
Related resources
- documented by: MobileWorld: Benchmarking Autonomous Mobile Agents in Agent-User Interactive and MCP-Augmented Environments
Primary evidence and verification
Source retrieval is separate from experiment reproduction. Use the original publication or publisher page when citing claims about the resource.
- editorial primary review
Retrieved 2026-10-17T09:12:16Z. Supports: name, kind, summary, environments, topics, scope, notes, relations, details.
Content fingerprint
fbc306b3414dfdb230171358efd6eb8d67573fbb94e11487a4a7e971e2abf0d9 - editorial primary review
Retrieved 2026-10-17T09:12:15Z. Supports: identity.
Content fingerprint
266e1fe5b5b1a4c1540ab0fd5eba206395084ee283f0079f424930a156cf8e18 - editorial primary review
Retrieved 2026-10-17T09:14:08Z. Supports: release_status.
Content fingerprint
14500ed2dfa14bb5c34211090f0c8827e4a16c3233e35a2d695c0d7261f9358b - editorial primary review
Retrieved 2026-10-17T09:14:09Z. Supports: licenses.code.
Content fingerprint
b09fcaf4fbc265223c2479c410beef2b98c5e687c78ecedf3452aa12ffc465f1
Suggest a correction
Open a source-backed correction. Include this record ID and the original source. Read our editorial policy and coverage boundaries.