Google Cloud Architecture Framework: Security, privacy, and compliance: Major update to align the recommendations with core principles of security.
New Dataproc on Compute Engine subminor image versions:
Google NotebookLM Enterprise: Add content from Agentspace Enterprise to notebooks (GA)
If you use Agentspace Enterprise, you can add some kinds of results, such as Google Sheets and Google Docs, from Agentspace Enterprise searches to your notebooks. Notebooks that contain sources from Agentspace Enterprise can't be shared. Also, shared notebooks can't accept sources from Agentspace Enterprise.
For more information, see Use NotebookLM Enterprise within Agentspace Enterprise and Notebooks that can't be shared.
Generally available: Support for service account JWT authentication for Identity Platform and Workforce Identity Federation configured applications. For more information, see Authenticating with a service account JWT.
deepseek-ai/deepseek-r1 and microsoft/Phi-4 models were added to Model Garden.
Generally available: You can configure Workforce Identity Federation with IAP, and use an external identity provider (IdP) to authenticate and authorize a workforce—a group of users, such as employees, partners, and contractors—using Identity and Access Management (IAM), so that the users can securely access services deployed on Google Cloud or on-premises.
For more information, see Configure IAP with Workforce Identity Federation.
Starting March 2025, the default version for new Cloud Composer environments changes from Cloud Composer 2 to Cloud Composer 3.
New environments that are created without a specified version will use the default Airflow build of Cloud Composer 3, composer-3-airflow-2
. Currently, the default version is composer-2-airflow-2
.
Google Agentspace Enterprise: Integration of NotebookLM Enterprise (GA)
If you use Agentspace Enterprise, you can access your notebooks and all the features of NotebookLM Enterprise from within the Agentspace Enterprise web app UI by clicking the NotebookLM button.
For more information, see Use NotebookLM Enterprise within Agentspace Enterprise.
The following advanced LLM inference optimization techniques are available in Model Garden in Preview:
Prefix caching reuses computations from previously generated text, eliminating redundant processing. It reduces time-to-first-token for requests with common prompt prefixes. Prefix caching is available for the following models:
Speculative decoding is an effective optimization technique to reduce generation time-per-output-token latency. For more information, see the Model Garden advanced features notebook.
Support for retrieving the detailed history of a workflow execution (expected iterations, in-scope variables) is available in Preview.
Google Cloud Console has failed to load JavaScript sources from www.gstatic.com.
Possible reasons are: