Google's Android Bench adds new LLMs; Fable 5 leads, Gemini lags

Google has updated its Android Bench benchmark for evaluating large language models (LLMs) in Android development tasks. The updated leaderboard includes eight new models, such as Claude Fable 5, Claude Sonnet 5, and Qwen 3.7 Max. Notably, Claude Fable 5 leads in accuracy at 84.5 percent, while Google's own Gemini 3.1 Pro ranks fifth. The benchmark also highlights significant cost differences between models, with Fable 5 and GPT 5.5 being the most expensive to run. AI

IMPACT Provides developers with updated performance data to select the best LLMs for Android coding tasks.

RANK_REASON Update to an existing benchmark tool for LLMs in software development.

Read on Ars Technica — AI →

• Android Bench
• Claude Fable 5
• Claude Sonnet 5
• Gemini 3.1 Pro
• Gemini 3.5 Flash
• Google
• GPT 5.4
• GPT 5.5
• Qwen 3.7 Max

• product
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →