Releasesby GlyphGKE capacity buffers are generally available
Capacity buffers for GKE Standard and Autopilot are now generally available. This feature allows users to pre-provision and suspend compute capacity to accelerate workload scaling.
Also shipped Nov 11, 2026Google Cloud in November 2026 →
Web search in apps
Retool added native web search capabilities to the app builder Chat tab and app functions, enabling agent applications to query live web data via a managed Tavily account.
Eval datasets, curated by Bayes from your agents' real conversations
Monte Carlo introduced automated evaluation dataset generation by Bayes, creating Golden and Improvement datasets from an agent's scored conversations across intent and issue clusters.
Pages are generally available
Databricks announced the general availability of Pages in Unity Catalog, providing governed definitions of business concepts organized by domain on the Discover page.
Ornith-1.0-9B and Ornith-1.0-35B models now available on Amazon SageMaker JumpStart
DeepReinforce AI's Ornith-1.0-9B and Ornith-1.0-35B models are now available on Amazon SageMaker JumpStart. These open-source models are designed for agentic coding workflows.
DeepSeek-V4-Pro, DeepSeek-V4-Flash, and DeepSeek-V4.1-Flash models now available on Amazon SageMaker JumpStart
DeepSeek-V4-Pro, DeepSeek-V4-Flash, and DeepSeek-V4.1-Flash models are now available on Amazon SageMaker JumpStart. These mixture-of-experts models are designed for reasoning, coding, and agentic workloads.
Issue types and fields are available for personal accounts
Issue types and issue fields are now generally available for repositories owned by personal accounts. This allows individual developers to standardize workflows and organize issues with custom fields and types.
More from Google Kubernetes EngineEvery Google Kubernetes Engine release →
Agent Sandbox on GKE is generally available and runs on any supported GKE version
Agent Sandbox on GKE is generally available and runs on any supported GKE version. Agent Sandbox on GKE is optimized for running isolated, stateful, single-replica workloads, such as AI agents, at no additional cost. You can install Agent Sandbox on GKE by using the open source Agent Sandbox installation .
Agent Gateway for GKE is now available in Preview in the Rapid channel on clusters running GKE version...
Agent Gateway for GKE is now available in Preview in the Rapid channel on clusters running GKE version 1.37.2-gke.1042000 or later. Agent Gateway provides a centralized management layer to control traffic flowing to, from, and between your agents on GKE.
In GKE version 1.35.1-gke.1516000 or later, GKE Dataplane V2 clusters use a unified cilium-cni configuration...
In GKE version 1.35.1-gke.1516000 or later, GKE Dataplane V2 clusters use a unified cilium-cni configuration instead of a chained CNI setup. With this change, Cilium configures the Pod default gateway by using its node-local virtual router interface ( cilium_host ), which has the statically assigned link-local IP address 169.254.4.6 .
(2026-R47) Version updates
GKE cluster versions have been updated. New versions available for upgrades and new clusters. The following versions are now available for new GKE clusters, and for manual control plane upgrades and node upgrades for existing clusters. For more information about versioning and upgrades, see GKE versioning and support and About GKE cluster upgrades .
Note : Your clusters might not have these versions available
Note : Your clusters might not have these versions available. Rollouts are already in progress when we publish the release notes, and can take multiple days to complete across all Google Cloud zones. Version 1.37.1-gke.2097000 is now the default version for cluster creation in the Rapid channel.
GKE Gateway adds SNI-based routing with TLS passthrough
GKE Gateway now supports TLS passthrough, allowing traffic routing based on SNI without terminating TLS at the load balancer.
Agent Substrate on GKE is generally available
Agent Substrate is now generally available for GKE Standard clusters. It provides an infrastructure layer optimized for high-density AI agent workloads with efficient sandbox provisioning.
In GKE version 1.37.0-gke.3570000 and later, you can split each Local SSD on a node between ephemeral storage...
In GKE version 1.37.0-gke.3570000 and later, you can split each Local SSD on a node between ephemeral storage and raw block storage. When you create a cluster or node pool, use the capacity property of the --ephemeral-storage-local-ssd flag to set how much Local SSD capacity GKE formats for ephemeral storage.