Google Updates Android Bench List

Google's updated Android Bench list shows that its own Gemini 4 Argon programming capabilities are significantly behind Claude and GPT. Gemini 4 Argon has a pass rate of 21.3% and a completion rate of 81.8%, ranking fifth, far behind Claude Sonnet 5.5 (pass rate 40%, completion rate 80.9% at the top) and GPT-6.1 Sol (pass rate 29.3%, completion rate 82.8% second), with an average call cost of 267.70 US dollars. Google announced the Android AI development benchmark Android Bench score, Kimi K3.1 Pro using kimi-code intelligence passed 24.0% in 30 tasks.

AI-written from the reports · Updated 4 days ago

1. 11 Nov, 08:56 · 1 reports
   Google updates Android AI development benchmark list
   ITHome · AI：Google updates Android AI development benchmark list: Kimi K3.1 Pro / DeepSeek V4.1 Flash scores revealed
2. 28 Oct, 10:09 · 1 reports
   Google Updates Android Bench List
   ITHome · AI：Google updates Android Bench list: its own Gemini 4 Argon programming ability significantly lags behind Claude and GPT