Google has updated its Android Bench leaderboard, a tool designed to evaluate AI models for Android development tasks. The benchmark assesses models based on specific Android development scenarios and best practices, while now incorporating average API call costs to assist developers in their selection process.
The latest update, released on October 26, includes performance results for several new models. Claude Sonnet 5.5 currently holds the top position with a 40% pass rate and an 80.9% completion rate. Claude Opus 5.5, added earlier in October, ranks second.
OpenAI's GPT-6.1 Sol model demonstrated a 29.3% pass rate and an 82.8% completion rate, with an average call cost of $99.50. In comparison, the top-ranked Claude model incurs a significantly higher cost of $1,574.90 per call.
Google's own Gemini 4 Argon model debuted in fifth place, trailing behind competitors and even the Claude Fable 5.1 model. Gemini 4 Argon recorded a 21.3% pass rate and an 81.8% completion rate, with an average call cost of $267.70.
