What's new in ML Runtimes version 2026.04.1

Review the major features and updates for ML Runtimes.

New Features/Improvements

New Runtimes:
- Python 3.14 PBJ Workbench - Hardened Edition
- Python 3.14 JupyterLab - Hardened Edition
- Python 3.11 PBJ Workbench - Hardened Edition
- Python 3.11 JupyterLab - Hardened Edition
Hardened Chainguard Runtimes: This release introduces new Hardened Edition ML Runtimes based on Chainguard images. These runtimes are designed to meet strict security standards and provide enhanced protection for your workloads and are released behind a paywall.
note
Hardened Edition Runtimes will be available in the upcoming Cloudera AI Public Cloud and On Premise releases.
Python 3.14 minor version has been added.
Existing Python maintenance versions have been upgraded to:
- Python 3.10.20 (from 3.10.19)
- Python 3.11.15 (from 3.11.14)
- Python 3.12.13 (from 3.12.12)
- Python 3.13.12 (from 3.13.9)
The existing R maintenance version has been upgraded to R 4.5.2 from 4.5.0.
Cloudera Copilot now supports the latest Bedrock models including Claude 4.x and the Llama 4 family across all JupyterLab runtimes.
To ensure stricter data residency and compliance, the service has transitioned from global inference profiles to region-specific model identifiers. This update ensures that all Copilot traffic remains securely within your configured AWS region while providing access to advanced generative AI capabilities.

JupyterLab Executable Path Override: There is a breaking change affecting all JupyterLab Runtimes in this release. If you use the JUPYTERLAB_DIR_OVERRIDE environment variable, you must update your configuration to point directly to the jupyter executable file.
Previously, the runtime engine looked for the executable in a /bin subdirectory (${JUPYTERLAB_DIR_OVERRIDE}/bin), but it now requires the exact file path.
A new Edition, called Hardened, has been added to the Runtime Catalog page.