Google’s Android coding tests reveal an unexpected Gemini 3.5 Flash weakness

Credit: Joe Maring / Android Authority TL;DR Google’s Android Bench results show Gemini 3.5 Flash trailing older models despite its premium positioning. Gemini 3.5 Flash missed the top five, while OpenAI’s GPT 5.5 claimed first place and Gemini 3.1 Pro Preview outperformed its successor. Google’s newest Flash model scored 63.7 and became the most expensive option in the rankings, averaging $147.1 per run.