Google has recently updated its Android Bench tool for evaluating large language models (LLMs) in the field of Android app development. Now, at the top of this benchmark's leaderboard, new models such as Claude Fable 5, Claude Sonnet 5, Claude Opus 4.8, GLM 5.2, Kimi K2.7 Code, MiniMax M3, Qwen 3.7 Plus, and Qwen 3.7 Max are featured.
Launched since March, this tool has also added criteria such as cost and efficiency to its evaluations, and developers can run their own tests and provide feedback that will help improve the future of Android Bench. The goal of this update is to clarify the performance of AI models in completing one hundred different Android programming tasks and to help developers choose the most optimal tool.