Inkling Small launches: 276B parameter model runs on a single Mac Studio with MLX

pcuenq · x · 2026-07-31

Inkling Small has been released with 276B parameters but massively reduced deployment requirements. The nvfp4 checkpoint runs on a single Mac Studio with MLX, taking 150 GB, and can also run across two 128GB laptops.

Related event: Thinking Machines Releases Open-Source Inkling-Small Model(18 posts)→

Original post →

More from Models

Models channel →