Benchmark Almanac

SPA-Bench

M3A 34.7% tops its leaderboard CogAgent 1.35% leads the models we track
Not in model cards

Agents use Android apps such as Airbnb, Spotify, and WeChat. The grader matches screenshots against task requirements and uses a vision model to resolve uncertain cases.

OpenMobile-2 ran difficulty level 3 on physical phones. The full benchmark has 300 single-app tasks and 40 cross-app tasks across 66 apps.

Released
October 2024
Kind
Phones
Level
Everyday
Scored
Graded by a model
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English, Chinese
In and out
Text, Screens → Actions
Interaction
Agentic
Refresh
One fixed set
Run by
Anyone
Last model added
December 2023
Scores refreshed
November 20th, 2026
10
Entries in its paper, 1 model we track

M3A tops the leaderboard with 34.7% CogAgent leads with 1.35%

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 combined score, so this is the plain mean of single-app and cross-app overall success rates. 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 CogAgent, at 1.35%. This view shows only models we track, each at its best.

  1. 1 M3A 34.7%
  2. 2 MobileAgentV2 26.9%
  3. 3 T3A 26.7%
  4. 4 SeeAct 21.8%
  5. 5 MobileAgent 19.4%
  6. 6 AppAgent 14.7%
  7. 7 OdysseyAgent 1.85%
  8. 8 CogAgent 1.35%
  9. 9 Auto-UI 0.50%
  10. 10 DigiRL 0.50%

Source: The SPA-Bench paper

Source: The SPA-Bench paper

#EntryScoreSettingDateSource
1 M3A 34.7% Official leaderboard
2 MobileAgentV2 26.9% Official leaderboard
3 T3A 26.7% Official leaderboard
4 SeeAct 21.8% Official leaderboard
5 MobileAgent 19.4% Official leaderboard
6 AppAgent 14.7% Official leaderboard
7 OdysseyAgent 1.85% Official leaderboard
8 CogAgent 1.35% Dec 2023 Official leaderboard
9 Auto-UI 0.50% Official leaderboard
10 DigiRL 0.50% Official leaderboard
#ModelScoreSettingDateSource
1 CogAgent 1.35% 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 SPA-Bench 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 (10 entries).
  • Pass Models scored 1 model we track, plus 9 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 1.35%, 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 9 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.

How we check