Spend caps in the Gemini API, available starting today!! This is another step forward of many to give developers more control and peace of mind when building with Gemini. Please go set a cap and send us any feedback as you use them!

Set a spend cap: Here are a few things to note: - Spend caps can have up to a 10 minute delay before taking effect, you are still responsible for the overages in this case - We are working to bring this latency down over time - We are shortly rolling out email notifications when you hit caps - Spend caps are experimental so pls send us feedback on things you want here!

Great work by @jastephx and so many others across Google to help land this!!

Share your API key so we can test :)

Per project and not per key? Useless, sorry. How can i have 200 projects? makes no sense to not control per key, every other AI Model company does...

not key for now, rate limits and spending all happen at the project level today

Yes, so sad... there is no control, imagine i have 10, 30 keys, cant control them, need to use proxy, ai gateways, really important to have that, is even in the roadmap?

finally i can sleep without checking the gemini billing dashboard every twenty minutes. you too?

yes

Great work by the Team!

Finally! Great work. Question: I am trying to understand what "experimental" and "preview" mean for models and other features. Does it mean that things aren't working as they should? Is there a timeframe before they become "production" ready?

@roeymiterany both of these just mean that we might not support the long term

@OfficialLoganK got it. Thanks for the clarification 🙏

nice, just heads up the caps apply per request not per minute so batch calls can still hit limits unexpectedly

Is this feature just for AI Studio? I can't seem to find the spending cap in the Google Cloud Console.

Finally. After I saw that guy hitting $86k in unauthorized charges with the API, I was paranoid

Amazing work shipping this! Real quality of life improvement for paying users.

Spend caps in the Gemini API solve the single biggest anxiety every developer has when building with AI: the runaway bill. One misconfigured loop, one agent that goes rogue, one weekend where you're not monitoring — and you come back to a bill that exceeds your entire project budget. Spend caps turn "I hope nothing goes wrong" into "the worst case is defined." That peace of mind is worth more than any feature improvement. For agentic workflows — where AI agents run autonomously for hours, making API calls without human oversight — spend caps aren't a nice-to-have. They're a prerequisite for deployment. No production system runs agents without cost guardrails. No CTO approves agent infrastructure without a hard ceiling on spend. This feature doesn't just help individual developers. It unblocks entire categories of autonomous agent deployment that were too financially risky without it. Email notifications when you hit caps — rolling out shortly — completes the control loop. Cap sets the ceiling. Notification tells you when you've reached it. Together they give developers both the safety net and the awareness to manage costs proactively rather than reactively. The developer who gets an email at 80% of their cap can investigate and adjust. The developer who discovers the overage on their monthly bill can only regret.

Definitely needed this, thanks a lot

Cost visibility is something developers have been asking for. Good step.

Gemini is so rate limited right now, there is no need for stinking spend cap 🤣

Thank you 🙏 so much! 🧢

Spend caps lower the risk floor for devs testing Gemini in production.

nice! it would be great if we could set the date filter to "all time". highest rn is 90 days.

cc @jastephx

Great. Now fix your damn Antigravity product. As a Pro user I was waiting the WHOLE week for my limits to be refreshed. When the day arrived - I was out of quota again WITHOUT EVEN USING ANY PROMPTS. Now it says I need to wait until next week again.

very nice, implemented immediately!

Thank you! This was much needed!

finally 🥳

Needed this. Was installing Claude api before I realized Gemini api was more efficient

This should be default, not opt-in. Seen too many $5K surprise bills from rogue loops.

Finally! 🔥

Peace of mind for the CFO. Infinite debt for the developer who forgets to scale.

This is the most needed request of all time !!!

Pro/Ultra subscription integration Ai Studio

nothing glamorous here, just a very real reason more teams will feel safe shipping on gemini.

Thanks ! This is very helpful to those of us deploying production applications. Is it possible to set the cap via API?

not right now, only in the UI

Got it. Thanks.

Can we apply this also to the vertex api ?

Gemini3 processing in UK please. Stuck using 2.5 for data residency reasons

Awesome! And vertex?

Great news!

Very important since API costs can add up quickly especially when using AI agents.