Hugging Face Hires oMLX Creator Jun Kim to Accelerate Apple MLX
Hugging Face hires oMLX creator Jun Kim to fund faster MLX development, keep Apache 2.0 licensing and speed Transformers support on Apple Silicon devices.
Summary
Hugging Face hired oMLX creator and maintainer Jun Kim on September 22, 2026, turning his side project into funded, fully maintained work. Kim will continue leading oMLX and guiding contributors, while the project retains its Apache 2.0 license. oMLX supports MLX, Apple’s local AI framework optimized for Apple Silicon. Hugging Face has backed MLX since Awni and Angelos introduced it in 2023, and its Hub hosts community MLX models.
Hugging Face plans to use oMLX as a testbed while building on mlx-lm and mlx-vlm, upstreaming useful work and deepening collaboration with LMStudio and the teams of Cheng, Prince and Yagil. Its priority is rapidly converting Hugging Face Transformers model definitions into reference MLX implementations usable by multiple engines, allowing each to focus on distinct features and bringing new models to MLX faster. An immediate access issue remains: commenters Alban Seurat and Alexey Satonin said the oMLX GitHub repository appeared deleted and a DMG was unavailable, with no replacement URL provided.
Positives
- Jun Kim’s move to Hugging Face gives oMLX dedicated funding and full maintenance instead of side-project status.
- Apache 2.0 licensing will remain in place, preserving open access to oMLX.
- Reference MLX implementations could help multiple engines adopt new Hugging Face Transformers models faster.
- Collaboration with mlx-lm, mlx-vlm, LMStudio and other teams could broaden shared improvements across the MLX ecosystem.
- Hugging Face’s Hub already gives users a central place to find and contribute MLX models.
Risks & concerns
- Alban Seurat said the oMLX GitHub repository appeared to have been deleted, leaving its current location unclear.
- Alexey Satonin reported being unable to obtain a DMG while the repository was unavailable.
- No replacement repository URL was provided in the supplied discussion.
- oMLX depends on foundational projects including mlx-lm and mlx-vlm, making coordination and upstream integration important to its plans.