Thinking Machines Lab Releases Inkling 2 Mini for Local Devices
Unite.AI

Thinking Machines Lab launched Inkling 2 Mini, a 26.4B-parameter open-weights model designed for local use on consumer laptops with about 20GB memory. It features a novel looping Mixture-of-Experts transformer architecture to enhance reasoning and is available on Hugging Face with multiple local inference options.
Thinking Machines Lab launches Inkling 2 Mini with 1M-token context for local AISee full coverageFull coverage from 2 sources
