Google's new Gemini 3.5 Flash-Lite model is more intelligent than Gemini 3 and outperforms Gemini 3.1 Flash-Lite across most use cases, while maintaining the same cost as previous versions. The release also signals the approaching end-of-life for the Gemini 2.5 Flash model.
Google is driving down the cost-performance ratio for its smaller, faster models, increasing the efficiency of inference compute for its cloud customers.