Gemini 3.7 Flash just dropped and the economics are insane:

20% fewer tokens than 3.6
50% cheaper per token
= 2.5x cost reduction end-to-end

And the output quality actually improved.

This isn't an incremental upgrade. This changes the entire AI cost structure for builders. If you're still pricing products around GPT-4 economics, you're about to get undercut hard.

Google just made AI apps 2.5x more profitable overnight.