Gemini 4 Argon ranks fifth in Android coding benchmark
Claude Sonnet 5.5 led Google's Android Bench with a 40% pass rate, while an October 26 update added reasoning-level details.
Google's October 26 update to Android Bench placed its Gemini 4 Argon model fifth, with a pass rate of 21.3%. Claude Sonnet 5.5 took the top position in the company's assessment of AI assistance for Android programming.
The benchmark gave Claude Sonnet 5.5 a 40% pass rate and an 80.9% completion rate. Claude Opus 5.5 held second place after being added in early October. Gemini 4 Argon ranked below Claude Fable 5.1 and recorded a completion rate of 81.8%.
OpenAI's GPT-6.1 Sol registered a 29.3% pass rate, with completion at 82.8%. The latest additions covered that model, GPT-6 Sol, Claude Sonnet 5, Claude Sonnet 5.5 and Gemini 4 Argon. Google also began specifying the reasoning level used in testing each model.
Introduced earlier in the year, Android Bench tests models against requirements specific to Android programming and development practices. Google has expanded the rankings and added average costs as its testing has continued.
