Mobile-Bench-v2 released data

Offline mobile-agent task data spanning common tasks and released noisy, contaminated and ambiguous-instruction subsets.

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, robustness

Record publication date

Not recorded

Evidence retrieved

2026-10-17T04:01:33Z

Resource ID

url:1d5b85321b98765141fc

What licenses are recorded?

Code, training data, weights, and documents can have different terms. A blank declaration means unknown, not unrestricted use.

Code

Not recorded

Data

Not recorded

Document

Not recorded

Weights

Not recorded

License review notes

• Data: The same first-party card has an Apache-2.0 metadata header but explicitly assigns Apache-2.0 to code and CC-BY-NC-SA-4.0 to the dataset in its License section. Preserve both declarations and leave a single data license unresolved instead of displaying the permissive header as the data permission. Source 1

Evaluation protocol

Focus

offline multi-path evaluation, noisy interfaces, ambiguous instructions

Setting

offline

Observations

mobile screenshots, XML UI state, instructions

Actions / outputs

mobile UI actions, clarification responses

Evaluation

Paper defines multi-path and noise/ambiguity evaluations; the public card identifies which files are full common-task sets and which are samples.

Scale and counting units

Release completeness differs by split. No full-benchmark sample total is inferred from the archive manifest.

Compare source-reviewed profiles

Recorded specifications

disabled

false

download url

https://huggingface.co/datasets/xwk123/MobileBench-v2/tree/main

gated

false

hub id

xwk123/MobileBench-v2

Explore related topics

Scope and limitations

The card distinguishes complete common-simple/common-complex JSON files from sampled noisy and ambiguous subsets; common tasks also require the separately cataloged Mobile3M dataset. Card metadata says Apache-2.0, while the prose explicitly assigns Apache-2.0 to code and CC-BY-NC-SA-4.0 to data. Do not apply the permissive header to dataset payloads, and do not describe every announced split as fully released.

Related resources

• documented by: Mobile-Bench-v2: A More Realistic and Comprehensive Benchmark for VLM-based Mobile Agents

Primary evidence and verification

Source retrieval is separate from experiment reproduction. Use the original publication or publisher page when citing claims about the resource.

1. editorial primary review

   Retrieved 2026-10-17T04:01:33Z. Supports: name, kind, summary, topics, scope, relations, details, licenses.data.

   Content fingerprint

   13424be795c7b75b8c2195469c1252a7b3ec633e8bb31f1f6d7e44476611cbe9
2. editorial primary review

   Retrieved 2026-10-17T04:01:33Z. Supports: identity, release_status, details.

   Content fingerprint

   247b1766b8cbdb2943c4602d2667f9f06dec8b84f0d24457080d499177475f22
3. editorial primary review

   Retrieved 2026-10-16T06:31:42Z. Supports: relations, summary.

   Content fingerprint

   afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e
4. editorial primary review

   Retrieved 2026-10-16T06:38:26Z. Supports: environments, notes.

   Content fingerprint

   356a8dca28fa3dc46bc05b13eaf04a3db273e63c06be4fd15d0e87e0343c0824

Suggest a correction

Open a source-backed correction. Include this record ID and the original source. Read our editorial policy and coverage boundaries.