I m growing a small SaaS. And cloud costs are starting to hurt. I keep hearing about founders stacking $100-300k in Google Cloud credits, but all the advice feels vague or locked behind big-name accelerators.
Where did you actually get credits?
Any creative hacks or things to avoid?
If you ve cracked this, I d love to hear what worked.
And if you re still figuring it out too, just drop a comment. If I ve gathered some useful stuff, I'll be happy to share.
Google just introduced Gemini Decision 1
A new model built for fast, structured decisions through the Gemini API.
Handles up to 1,000 questions in a single request
Supports text, images, audio, video, and PDFs
256K context window
Returns decisions with confidence scores in around 150–250 ms
Costs just $0.10 per 1M input tokens, with no output-token charges
Free API tier available for testing
Great for AI agents, model routing, classification, and automated evaluations.
Try it in the Gemini API!
The confidence scores are the part that caught my eye — I've been wiring up a lot of small either/or decisions in my agent work and usually end up hand-rolling thresholds because raw model outputs aren't well calibrated. At 150–250ms with up to 1,000 questions batched, this looks like it could replace a bunch of that glue. Curious how the confidence holds up on ambiguous or borderline inputs — is it something you'd trust to auto-route, or more of a signal to flag for a second look?