AndroidWorld
We show this table for reference; we do not rank on it.
A mobile GUI agent benchmark for completing Android app workflows and on-device tasks.
Benchmark score on AndroidWorld — November 11, 2026
We compile the AndroidWorld rows from provider self-reports and secondary reports. Qwen3.8-Omni-Flash leads the table at 87.1%, followed by Qwen3.8 Max (85.3%) and Qwen3.8-Flash-Next (84.5%). We do not use these results to rank models overall.
Qwen3.8-Omni-Flash
Alibaba
Qwen3.8 Max
Alibaba
Qwen3.8-Flash-Next
Alibaba
10 modelsAgenticCurrentDisplay onlyUpdated November 11, 2026
Benchmark score table (10 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Qwen3.8-Omni-FlashAlibaba | 87.1% | Not reported | Closed |
| 2 | Qwen3.8 MaxAlibaba | 85.3% | Not reported | Open |
| 3 | Qwen3.8-Flash-NextAlibaba | 84.5% | Not reported | Open |
| 4 | Qwen3.8-27BAlibaba | 81.9% | Not reported | Open |
| 5 | Qwen3.7 PlusAlibaba | 81.0% | Not reported | Closed |
| 6 | Holo3.1-35B-A3BH Company | 79.3% | Not reported | Open |
| 7 | Holo3.1-4BH Company | 71.0% | Not reported | Open |
| 8 | Holo3.1-9BH Company | 71.0% | Not reported | Open |
| 9 | Qwen3.6-27BAlibaba | 70.3% | Not reported | Open |
| 10 | Claude Opus 4.6Anthropic | 62.0% | Not reported | Closed |
Among the reported AndroidWorld rows, Qwen3.8-Omni-Flash is first at 87.1%. The third row is 2.6 points behind. The broader top-10 range is 25.1 points, so the table still separates the published systems.
10 models have been evaluated on AndroidWorld. The benchmark falls in the Agentic category. AndroidWorld is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About AndroidWorld
Year
2026
Tasks
Android app workflows
Format
Interactive mobile-agent evaluation
Difficulty
Complex mobile task completion
BenchLM tracks AndroidWorld as a display-only benchmark reference when providers publish exact values alongside broader GUI-agent summaries.
Freshness and provenance
Version
AndroidWorld 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does AndroidWorld measure?
A mobile GUI agent benchmark for completing Android app workflows and on-device tasks.
Which model scores highest on AndroidWorld?
Qwen3.8-Omni-Flash by Alibaba currently leads with a score of 87.1% on AndroidWorld.
How many models are evaluated on AndroidWorld?
10 AI models have published results on AndroidWorld in the BenchLM catalog.
Compare top models on AndroidWorld
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.