Here we are post 3X announcement.
This was after a productive hour or refactoring.
The previous flash model was so efficient, I would have had 70% left of quota easily.
If I had the Flash 3.0 I would have continued refactoring.
We are going to miss the golden days of Flash 3.0.
Google stated Flash models were supposed to be the “use without thinking twice” option; fast, lightweight, and quota friendly.
One simple refactor cost me 20% of Gemini 3.5 Flash and 100% of Opus quota, and the task still wasn’t fully completed. That honestly says everything.
Flash 3.0 may not have been the smartest model on paper, but it absolutely nailed the real-world developer experience. That balance is missing now.
