Thinking MachinesLaunchAI ModelsNovember 18, 2026

Thinking Machines releases Inkling 2 Mini, a 26B looped transformer for local use

Read original source →huggingface.co

Inkling 2 Mini is a multimodal MoE with ~26B total and 3B active parameters per token; its 12 middle layers run three times, giving 48 logical layer visits from 24 physical ones. It supports 1M-token context and ships Apache 2.0 with NVFP4/MXFP8 weights and a DSpark drafter.

How this story unfolded

1 day · 5 reports · 34 community posts · 39 of 44 shown

  1. Nov 20
  2. Nov 21

More stories today

Open the live feed