Download Latest Version onnxruntime-win-x64-gpu_cuda13-1.28.2.zip (365.6 MB)
Email in envelope

Get an email when there's a new version of ONNX Runtime

Home / v1.28.1
Name Modified Size InfoDownloads / Week
Parent folder
onnxruntime-linux-x64-1.28.1.tgz 2026-08-18 9.1 MB
onnxruntime-linux-aarch64-1.28.1.tgz 2026-08-18 8.1 MB
onnxruntime-win-x64-gpu_cuda13-1.28.1.zip 2026-08-18 365.9 MB
onnxruntime-linux-x64-gpu_cuda13-1.28.1.tgz 2026-08-18 240.9 MB
onnxruntime-win-x64-gpu_cuda12-1.28.1.zip 2026-08-18 455.0 MB
onnxruntime-win-x64-1.28.1.zip 2026-08-18 78.6 MB
onnxruntime-win-arm64x-1.28.1.zip 2026-08-18 122.9 MB
onnxruntime-win-arm64-1.28.1.zip 2026-08-18 79.8 MB
onnxruntime-osx-arm64-1.28.1.tgz 2026-08-18 31.5 MB
onnxruntime-linux-x64-gpu_cuda12-1.28.1.tgz 2026-08-18 423.7 MB
ONNX Runtime v1.28.1 source code.tar.gz 2026-08-17 278.7 MB
ONNX Runtime v1.28.1 source code.zip 2026-08-17 285.6 MB
README.md 2026-08-17 1.6 kB
Totals: 13 Items   2.4 GB 19

This is a patch release on top of v1.28.0, containing support for device-free WebGPU compilation, improved compatibility with sandboxed Windows processes, and targeted graph-validation fixes.

WebGPU EP

  • Added support for device-free compile-only sessions, enabling offline graph transformation and optimized-model serialization without access to GPU hardware (#29681)

Bug Fixes

  • Prevented an access violation in Windows processes under Win32k lockdown by skipping DXGI device discovery (#29755)
  • Allowed zero-input EPContext nodes, aligning their schema with support for compiling zero-input models (#29799)
  • Hardened FastGelu fusion to skip malformed Mul and Pow patterns (#32016)
  • Added validation for in-memory external initializer references, rejecting unregistered or mismatched data before graph transformation (#32042)

Contributors

Thanks to our 4 contributors for this release!

@apsonawane, @shiyi9801, @adrastogi, @mingmingtasd

Full Changelog: v1.28.0...v1.28.1

Source: README.md, updated 2026-08-17