Cluster Toolkit 1.92.0: ML Diagnostics on TPUs, GKE Auto-Provisioning
Cluster Toolkit version 1.92.0 now supports ML Diagnostics on TPU machine types and node auto-provisioning for GKE clusters. This release also enables an optional infrastructure setup for inference gateways and compact placement for Slurm clusters via DWS flex-start provisioning. These updates benefit users working with machine learning on Google Cloud and managing GKE or Slurm environments.
Features (1) ›
- Cluster Toolkit
Cluster Toolkit version 1.92.0 is available. This release introduces support for ML Diagnostics on TPU machine types and node auto-provisioning on GKE clusters. The release also adds an optional infrastructure setup for inference gateways and compact placement for Slurm clusters by using the Dynamic Workload Scheduler (DWS) flex-start provisioning model. For details, see the Release announcement on GitHub .
https://docs.cloud.google.com/release-notes#June_04_2026
Related releases
- Cloud Load Balancing: Global External Passthrough Network Load Balancer Preview Google Cloud release notes ·
- Config Connector 1.154.1 Adds New Alpha Resources and Field Support Google Cloud release notes ·
- Google Cloud AI Infrastructure and Orchestration Updates - July 2026 Google Cloud Blog ·
- Google Kubernetes Engine: New versions and deprecations Google Cloud release notes ·
- Managed Service for Apache Spark drops default Conda channels Google Cloud release notes ·
- GKE Optimizations Reduce AI Agent Compute Costs Up To 75% Google Cloud Blog ·