AndroidWorld
AndroidWorld - visual agent benchmark for Android device task completion.
Category
General Agents
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for AndroidWorld
| Model | Score | Date | Verified |
|---|---|---|---|
| Qwen3.6-27B |
45.17%
|
22.04.2026 | Verified |
| Qwen3.5-27B |
21.62%
|
22.04.2026 | Unverified |
| Qwen3.8-27B |
89.96%
|
18.08.2026 | Verified |
| Qwen3.7-Plus |
86.49%
|
18.08.2026 | Unverified |
| Muse Glimmer-30B |
28.19%
|
18.08.2026 | Unverified |
| Opus4.6 Max |
13.13%
|
18.08.2026 | Unverified |
| Qwen3.5-4B |
58.60
|
01.02.2026 | Verified |
| Qwen3.5-122B-A10B |
30.12%
|
01.02.2026 | Verified |
| Qwen3.5-35B-A3B |
48.26%
|
01.02.2026 | Unverified |
| Qwen3-235B-A22B |
19.69%
|
01.02.2026 | Unverified |
| Qwen3-VL-235B-A22B |
19.69%
|
01.02.2026 | Verified |
| Qwen3.8-Flash-Next |
100.00%
|
26.08.2026 | Verified |
| Claude-Opus-4.6 (Max) |
13.13%
|
26.08.2026 | Unverified |