Public preview adds per-key monthly spend alerts and optional automatic pausing at 100% of budget

Paused keys resume at the start of the next billing cycle, or sooner if you raise the budget, turn off pausing, or delete the alert.

You can also have the alert pause the key when its spend reaches 100% of its budget, which stops Serverless Inference requests for that key without affecting other keys.

In public preview, you can now create spend alerts for Serverless Inference model access keys.