Jun Kim, the creator and maintainer of oMLX, has joined Hugging Face to support the MLX community, Hugging Face announced on September 22, 2026. The company described MLX, Apple's framework for local artificial intelligence optimized for Apple Silicon, as a central component of the local AI ecosystem.
The appointment transitions oMLX from a side project into a fully funded and maintained initiative. Hugging Face said that oMLX will retain its Apache 2.0 open-source license, with Kim continuing to direct development. Full-time backing will allow Kim to guide external contributors and build the software for the long term.
Apple researchers Awni Hannun and Angelos Katharopoulos originally introduced MLX in December 2023. Hugging Face serves as a central hub where developers find and share MLX models. Under the new arrangement, Hugging Face expects oMLX to serve as a testbed for new concepts while relying on foundational dependencies such as mlx-lm and mlx-vlm.
Hugging Face detailed technical plans to speed up how new model architectures reach MLX. The team plans to streamline the transition from transformers library definitions into reference MLX implementations that external inference engines can run, enabling different runtimes to focus on distinct engine-specific features.
The company also confirmed active collaborations with projects including LMStudio, mlx-lm, and mlx-vlm. Hugging Face stated that it plans to upstream relevant work and strengthen relationships with contributors Cheng, Prince, Yagil, and their development teams.
