Google's Gemini 3.5 Flash Lags in Android Benchmarks, Priciest Option
Google's refreshed Android Bench rankings reveal that its new Gemini 3.5 Flash model underperformed expectations in Android development tasks. Despite being the most expensive option at an average of $147.1 per run, Gemini 3.5 Flash did not make it into the top five. It scored 63.7, trailing behind older models, including its predecessor, Gemini 3.1 Pro Preview, and competitors like OpenAI's GPT 5.5.
Google's refreshed Android Bench rankings indicate that the newly introduced Gemini 3.5 Flash model did not meet performance expectations in Android development tasks. The model failed to secure a position within the top five, scoring 63.7. This benchmark evaluates the efficacy of various AI models in performing coding tasks relevant to Android development.
Despite its premium positioning, Gemini 3.5 Flash was identified as the most expensive option among those ranked, averaging $147.1 per run. It was noted to be three times the price of its predecessor, Gemini 3.1 Pro Preview, which achieved a score of 72.4.
The rankings show OpenAI's GPT 5.5 leading the list with a score of 74. Other high performers included GPT 5.4, which also scored 72.4, matching Gemini 3.1 Pro Preview. Additionally, various Claude Opus models were reported to have outperformed the Gemini 3.5 Flash variant in these tests.
(Source: Android Authority)
