Hugging Face hires Jun Kim to maintain oMLX
The local artificial intelligence project retains Jun Kim's leadership and its Apache 2.0 license, while maintenance becomes funded work.
oMLX maintenance becomes funded work. On September 22, 2026, Hugging Face announced that creator Jun Kim was joining the company to continue supporting the project and MLX, Apple's framework for local artificial intelligence.
The announcement says Kim keeps leading oMLX. Its Apache 2.0 license remains, and maintenance becomes funded work. Faster development is an expectation in the announcement, rather than an outcome it demonstrates.
The technical direction is more specific than the hiring news. Hugging Face wants to ease the transition from Transformers model definitions to reference MLX implementations that different inference engines can consume. It also describes oMLX as a place to explore ideas while contributing improvements to dependencies such as mlx-lm and mlx-vlm.
The funded role supports continuity, but does not justify replacing a working installation. Existing oMLX users and people evaluating local tools for Apple Silicon still need support for their particular model and workflow; the staffing announcement does not establish that compatibility.
Source: the Jun Kim and oMLX announcement, by Pedro Cuenca and coauthors · Hugging Face.