AI vendors and your data

Vendors146Answers427

146 vendors · 427 answers

Gemini Enterprise Agent Platform and zero data retention changed Google, Nov 9, 2026

Google · Nov 9, 2026 · 14 added, 4 removed · seen by the weekly check

The pageAll Google terms
  1. Added: Run the following curl command to turn off all caching for a Google Cloud project. This disables both in-memory caching and durable caching, and no cached data is retained. To run this command, a user must be granted the Agent Platform Administrator role, roles/aiplatform.admin .
  2. Removed: Run the following curl command to turn off all caching for a Google Cloud project. This disables both in-memory caching and durable caching, and no cached data is retained. To run this command, a user must be granted the Gemini Enterprise Agent Platform administrator role, roles/aiplatform.admin .
  3. Added: If you disabled caching for a Google Cloud project and want re-enable it, run the following curl command. To run this command, a user must be granted the Agent Platform Administrator role, roles/aiplatform.admin .
  4. Added: Run the following curl command to turn off durable caching and keep in-memory caching for a Google Cloud project. To run this command, a user must be granted the Agent Platform Administrator role, roles/aiplatform.admin .
  5. Added: If you turned off durable caching for a Google Cloud project and want to turn it back on, run the following curl command. To run this command, a user must be granted the Agent Platform Administrator role, roles/aiplatform.admin .
  6. Added: You can request zero data retention for supported partner models on Agent Platform.
  7. Added: Zero Data Retention (ZDR) is available for OpenAI flagship models on an approved project basis, but is disabled by default. Flagship models of the OpenAI GPT Astra, Sol, Terra, and Luna families are designated as Advanced AI under the Advanced AI Safety Addendum . For more information about safety and abuse monitoring for these models, see Non-Google Advanced AI models .
  8. Added: The following points summarize the default logging behavior and how ZDR operates when enabled:
  9. Added: Default state (ZDR off): ZDR is off by default for OpenAI flagship models and is configured at the Google Cloud project level. Existing ZDR approvals for the direct OpenAI API don't apply to Agent Platform, and setting store: false in Responses API requests doesn't enable ZDR on Agent Platform. By default, Google Cloud logs prompts, file inputs, and responses for up to 30 days in Google-owned storage in the us multi-region for safety and abuse monitoring. Customer data is never used to train or fine-tune models.
  10. Added: How ZDR works when approved: When ZDR is enabled on an approved Google Cloud project, Google Cloud processes requests in real time without writing prompts, file inputs, or responses to persistent disk. Customer prompts, file inputs, and model responses aren't stored in the 30-day abuse-monitoring logs. To reduce latency, Agent Platform uses volatile memory (RAM-only) prompt caching, where cached data is automatically evicted within 24 hours. Google Cloud continues to record operational and billing metadata (such as token counts and latency telemetry).
  11. Added: To request and verify ZDR for OpenAI flagship models, follow these steps:
  12. Added: Contact your Google Cloud account team and provide the Google Cloud project IDs and numbers where you require ZDR.
  13. Added: Your account team submits the ZDR request for OpenAI review and approval. After OpenAI approves the request, Google Cloud applies the ZDR configuration to your approved projects, which distributes to all serving infrastructure within 24 hours.
  14. Added: In the Google Cloud console, open Model Garden and navigate to the model page for an enabled OpenAI model in your approved project to verify that ZDR is active.
  15. Added: Check for the Zero Data Retention active confirmation banner on the model card. In an active ZDR project, 30-day prompt, file input, and response safety logging is disabled and acceptance of the Advanced AI Safety Addendum is no longer required.
  16. Removed: If you disabled caching for a Google Cloud project and want re-enable it, run the following curl command. To run this command, a user must be granted the Gemini Enterprise Agent Platform administrator role, roles/aiplatform.admin .
  17. Removed: Run the following curl command to turn off durable caching and keep in-memory caching for a Google Cloud project. To run this command, a user must be granted the Gemini Enterprise Agent Platform administrator role, roles/aiplatform.admin .
  18. Removed: If you turned off durable caching for a Google Cloud project and want to turn it back on, run the following curl command. To run this command, a user must be granted the Gemini Enterprise Agent Platform administrator role, roles/aiplatform.admin .
About this change
Page
docs.cloud.google.com/gemini-enterprise-agent-platform/resources/zero-data-retention
Kind
Documentation
Text hash
50723050aa0b to c2a36cb432cc
Dated by
the weekly check that first read the new text