Thinking Machines' small open-weights model that can run on a Mac: 26B parameters, 5.4B active, with looped layers for more computation per token.
594B Mixture-of-Experts model (17B active) with image input and up to 1M tokens of context, built for agentic coding and knowledge work.