Thinking Machines pushes open-weight models toward local deployment with Inkling 2 Mini
Thinking Machines Lab released Inkling 2 Mini, an Apache 2.0 open-weight 26B model using a looped-transformer design and targeting local use on consumer hardware. Coding and tool-use claims are company-reported; independent evaluation, reliability, and developer adoption will determine whether its efficiency proposition matters commercially.
Sources (4)
Updated Nov 22, 2026