IBM Power-Optimized Python Wheels

Edit online

IBM provides a curated repository of natively built Python wheels for IBM Power (ppc64le), optimized for AI, data science, and scientific computing. This document describes the Python wheels that are officially supported and validated for use with this release of the IBM Open-Source AI Foundation for Power.

Repository: https://wheels.developerfirst.ibm.com/ppc64le/linux-v2026.06.0

Supported platforms

Architecture: ppc64le
Processors: Power10, Power11
Python versions: 3.11, 3.12

Ecosystem coverage

The repository includes widely used AI/ML and data libraries such as:

PyTorch, TensorFlow (select versions)
NumPy, SciPy, pandas
scikit-learn
OpenBLAS

A complete package index is available in the PyEco repository:

Package index: https://github.com/ppc64le/pyeco/blob/v2026.06.0/DevpiWheelsIndex.md
Example AI/ML applications: https://github.com/ppc64le/pyeco/tree/v2026.06.0/examples

Note: IBM Open-Source AI Foundation for Power includes a selected set of supported Python wheels. To explore the full set of published Python wheels across additional Python versions and configurations, refer to the README.