onnxruntime-ep-mlx

MLX-native ONNX Runtime execution provider for Apple Silicon. It is an out-of-tree plugin EP for ONNX Runtime 1.29 and requires no ONNX Runtime fork.

The wheel bundles MLX 0.32.1 and mlx-c 0.6.0_4. It supports encoder, LLM prefill, and token-at-a-time decode workloads, including quantized matmul, attention, normalization, convolution, shape operations, and compiled shape-specialized If, Loop, and Scan control flow.

See the complete README for operator coverage, performance results, diagnostics, build instructions, and C/C++ integration.

Install

pip install -U onnxruntime-ep-mlx

import onnxruntime as ort
import onnxruntime_ep_mlx

onnxruntime_ep_mlx.register_execution_provider_library()
session = ort.InferenceSession(
    "model.onnx",
    providers=["MLXExecutionProvider", "CPUExecutionProvider"],
)
outputs = session.run(None, feeds)

The package requires macOS on Apple Silicon and an ONNX Runtime 1.29 build.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

onnxruntime_ep_mlx-0.29.6.dev1.tar.gz (291.7 kB view details)

Uploaded Sep 4, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

onnxruntime_ep_mlx-0.29.6.dev1-py3-none-macosx_14_0_arm64.whl (40.3 MB view details)

Uploaded Sep 4, 2026 Python 3macOS 14.0+ ARM64

File details

Details for the file onnxruntime_ep_mlx-0.29.6.dev1.tar.gz.

File metadata

Download URL: onnxruntime_ep_mlx-0.29.6.dev1.tar.gz
Upload date: Sep 4, 2026
Size: 291.7 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for onnxruntime_ep_mlx-0.29.6.dev1.tar.gz
Algorithm	Hash digest
SHA256	`66219a07bacc40f082579cd37431c8231bf4c46bbde0d55b6de293d05c8618d2`
MD5	`f2c29755a7e6cf940ed5c599436bfd07`
BLAKE2b-256	`b6c53835746f38802acf91f32ed975a100527efd291a9580a98e11b4e34d85d2`

See more details on using hashes here.

Provenance

The following attestation bundles were made for onnxruntime_ep_mlx-0.29.6.dev1.tar.gz:

Publisher: publish.yml on justinchuby/onnxruntime-mlx

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: onnxruntime_ep_mlx-0.29.6.dev1.tar.gz
- Subject digest: 66219a07bacc40f082579cd37431c8231bf4c46bbde0d55b6de293d05c8618d2
- Sigstore transparency entry: 2706818527
- Sigstore integration time: Sep 4, 2026
Source repository:
- Permalink: justinchuby/onnxruntime-mlx@5e45b30eee28161ce8bd1ca2de3faeaba0962162
- Branch / Tag: refs/tags/v0.29.6-dev.1
- Owner: https://github.com/justinchuby
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: publish.yml@5e45b30eee28161ce8bd1ca2de3faeaba0962162
- Trigger Event: release