Gemini releases have been disappointing for a very long time, but I'm genuinely intrigued by Gemini 3.7 Flash.
With Deepseek prices going up, Gemini 3.7 Flash is looking more attractive at $0.375/M input tokens $1.875/M output tokens.
The big benefit of this Google release its speed. It runs over 300 tokens per second, which makes iterations ridiculously fast. And this model's world knowledge is very impressive - much deeper than Deepseek V4 Flash. My initial tests show very good coding quality results.
This one will be fun to test because it's so outrageously quick. That speed makes a huge difference in wanting to use a model, because results are nearly instant.