Skip to content

Models Benchmarks Providers

Search models and benchmarks /

Android Control High_EM

reasoning

Categories: multimodal, reasoning
Modality: multimodal
Language: en
Multilingual: No
Max score: 1
Scoring: %, higher is better
Verified by llm-stats: No

Android device control benchmark using high exact match evaluation metric for assessing agent performance on mobile interface tasks

Leaderboard

Showing 3 of 3 results

Qwen2.5 VL 32B Instruct

69.6%

i
Qwen2.5 VL 72B Instruct

67.4%

i
Qwen2.5 VL 7B Instruct

60.1%

i

Wikibench About Theme Content licensed CC BY-SA 4.0.