Benchmark Almanac

MobileAgentBench

AppAgent + GPT-4V 40% tops its leaderboard CogAgent 8.00% leads as a model on its own
Not in model cards

Agents complete 100 tasks across ten Android apps. The grader checks app state and accessibility events, so agents can take different paths.

The V-Droid team used this benchmark as an external test and reported 49% task success on an Android emulator.

Released
June 2024
Kind
Phones
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Screens → Actions
Interaction
Agentic
Environment
Phone apps
Refresh
One fixed set
Run by
Anyone
In views
Computer use, Agents
Last model added
December 2023
Scores refreshed
November 20th, 2026
5
Entries in its paper, 1 model we track

AppAgent + GPT-4V tops the leaderboard with 40% CogAgent leads with 8.00%

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. The best model on its own is CogAgent, at 8.00%. Agents built around a model, such as AppAgent + GPT-4V, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

  1. 1 AppAgent + GPT-4V Android 14, self-exploration 40%
  2. 2 GPT-3.5 (snapshot unspecified) Android 9 27%
  3. 3 MobileAgent + GPT-4V Android 14 26%
  4. 4 AndroidArena + GPT-4V Android 14 22%
  5. 5 CogAgent Android 14, 4-bit 8.00%

Source: The MobileAgentBench paper

Source: The MobileAgentBench paper

#EntryScoreSettingDateSource
1 AppAgent + GPT-4V 40% Android 14, self-exploration Nov 2023 Official leaderboard
2 GPT-3.5 (snapshot unspecified) 27% Android 9, AutoDroid Official leaderboard
3 MobileAgent + GPT-4V 26% Android 14 Nov 2023 Official leaderboard
4 AndroidArena + GPT-4V 22% Android 14 Nov 2023 Official leaderboard
5 CogAgent 8.00% Android 14, 4-bit, CogAgent Dec 2023 Official leaderboard
#ModelScoreSettingDateSource
1 CogAgent 8.00% Android 14, 4-bit, CogAgent Dec 2023 Official leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

In the benchmark’s paper
0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) CogAgent 0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) CogAgent

Source: Official MobileAgentBench 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.

More in Computer and browser use

All 119 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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 2 years ago, in December 2023, for CogAgent. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (5 entries).
  • Pass Models scored 1 model we track, plus 4 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in December 2023, for CogAgent.
  • Note Frontier models None of its models are on the capabilities index.
  • Pass Headroom The best is 8.00%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on mobileagentbench.github.io: 4 rows for models we track and 1 other entry 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.

How we check