Veo 3.1 Fast: Billing and HTTP response codes when safety filters trigger

Google Cloud Build with AI Custom ML & MLOps
1

Hello,

I have questions regarding the billing behavior and HTTP responses for Veo 3.1 Fast when different safety filters are triggered.

I’ve reviewed the Responsible AI and usage guidelines here:

However, the documentation does not explicitly describe how billing and HTTP status codes are handled for different safety filter scenarios.

Could you please clarify the behavior for the following cases?

1. Prompt (input) safety block

When a prompt is blocked by safety filters before the model starts running:

  • What HTTP status code is returned?

  • Is the request billed or not?

  • Is there official documentation confirming the billing behavior in this scenario?

2. Output safety block

When Veo 3.1 Fast starts generation, but the output is blocked because the generated content is deemed unsafe:

  • What HTTP response code is returned?

  • Is the request billed?

  • Does billing depend on whether any frames were generated before the block?

3. Partial content blocked

When Veo 3.1 Fast generates partial content (for example, some frames or segments are filtered) and returns a partial or truncated video:

  • What HTTP status code is returned?

  • Is the request billed in this case?

  • Is billing proportional to generated content, or is the full request billed?

If possible, I would appreciate:

  • Official guidance on billing behavior for each scenario

  • Documentation links that explicitly confirm HTTP status codes and billing treatment

Thank you in advance for the clarification.