Android Agent Arena (201 tasks)
Version 201 tasks
Agents find live information and perform actions in Android apps such as YouTube Music and Booking.com. Task functions check app state and grade answers.
MobileJudgeBench ran 58 original tasks and released 154 trajectories. All 58 task IDs match the original repository, including two Booking.com tasks with date parameters.
- Made by
- Android Agent Arena team
- Released
- January 2025
- Kind
- Phones
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Phone apps
- Refresh
- New versions replace it
- Run by
- Anyone
- In views
- Computer use, Agents
- Last model added
- July 2024
- Scores refreshed
- November 20th, 2026
AppAgent tops the leaderboard with 13.3% InternVL2-8B leads with 10.3%
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The paper reports no overall score, so this is the plain mean of success rates at its three difficulty levels. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is InternVL2-8B, at 10.3%. This view shows only models we track, each at its best.
- 1 AppAgent 13.3%
- 2
InternVL2-8B Fine-tuned on AMEX and AndroidControl 10.3%
- 3 GPT-4o (snapshot unspecified) 3.77%
Source: The Android Agent Arena (201 tasks) paper
Source: The Android Agent Arena (201 tasks) paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | AppAgent | 13.3% | Official leaderboard | ||
| 2 | | 10.3% | Fine-tuned on AMEX and AndroidControl | Jul 2024 | Official leaderboard |
| 3 | GPT-4o (snapshot unspecified) | 3.77% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 10.3% | Fine-tuned on AMEX and AndroidControl | Jul 2024 | Official leaderboard |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Android Agent Arena (201 tasks) leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
Versions
More in Computer and browser use
- BrowseComp In use Hard-to-find facts that need persistent web browsing. Short verifiable answers graded for exact correctness. Claude Fable 5 93.3%
- ScreenSpot-Pro In use GUI grounding on high-resolution professional-software screenshots. The model must locate the target UI element. GPT-6 Astra 92.7%
- AndroidWorld In use Agents complete 116 tasks across 20 real Android apps in an emulator. Each task checks the device's system state to judge success. GPT-6 Astra 100%
- OSWorld 2.1 In use Agents complete 108 long-horizon workflows across 31 self-hosted websites and desktop apps, averaging 318 tool calls. Scored by binary completion and checkpoint partial credit. Claude Opus 5 44.3%
- WebArena-Verified In use Agents complete 812 audited WebArena tasks on self-hosted websites. Deterministic evaluators check the agent's answer and captured network traces.
Muse Spark 1.1 69%
- OSWorld 2.1 (offline subset) In use Agents complete the 82 OSWorld 2.1 tasks on the official offline list, with no internet access. Scored by checkpoint partial credit and strict pass rate. Claude Fable 5.2 89.3%
Data audit
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the benchmark’s paper (3 entries).
- Note Models scored Only 1 model we track, plus 2 other entries on its leaderboard.
- Note Newest score The newest came 2 years ago, in July 2024, for InternVL2-8B. That’s expected for a test replaced by Android Agent Arena (100 tasks).
- Note Frontier models None of its models are on the capabilities index.
- Pass Headroom The best is 10.3%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 1 row for models we track and 2 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.