Google Cloud announced the general availability of token-based cost management for generative AI functions in BigQuery on October 14, 2026. The new feature provides daily quotas to help administrators set self-serve guardrails on token consumption.
Daily Quotas and Controls
Google Cloud introduced four daily quotas managed through the Google Cloud Identity and Access Management quota system. These controls apply at both the project level and the per-user level.
Default values include 200 billion daily input tokens and 20 billion daily output tokens for a project, with per-user limits set to 150 billion input tokens and 15 billion output tokens. Output quotas also include reasoning or thinking tokens used by the large language model.
Covered Functions and Monitoring
The quotas apply to BigQuery LLM-backed generation functions including AI.GENERATE, AI.GENERATE_TABLE, AI.GENERATE_TEXT, AI.CLASSIFY, AI.IF, AI.SCORE, and AI.AGG. Embedding functions such as AI.GENERATE_EMBEDDING are not subject to these quotas.
Administrators can view historical token consumption, adjust limits in the Google Cloud console, or manage quotas programmatically using the Cloud Quotas API, gcloud CLI, and Terraform.
What to do
- Navigate to IAM & Admin and select Quotas & System Limits.
- Filter by service and select BigQuery API, then sort by the Name column.
- Find the quota, open the three-dot menu, and choose Edit quota to lower or increase limits.
Key facts and where they come from
- Google Cloud announced the general availability of token-based cost management for BigQuery generative AI functions.
We recently announced the general availability (GA) of token-based cost management for BigQuery generative AI functions
- Four daily quotas are managed through the Google Cloud Identity and Access Management quota system.
Four daily quotas, managed through the Google Cloud Identity and Access Management (IAM) quota system, provide controls at the project level and the per-user level.
- The default value for AI input tokens per day per project is 200 billion.
AI input tokens per day
Input tokens used by the LLM
Per day, per project
200 billion - Embedding functions are not subject to these quotas.
What's not covered: Embedding functions such as AI.GENERATE_EMBEDDING (also available as ML.GENERATE_EMBEDDING) and AI.EMBED are not subject to these quotas
