# Thinking Machines shrinks Inkling to 26B parameters and puts it on a laptop

> Source: <https://runtimewire.com/article/thinking-machines-shrinks-inkling-to-26b-parameters-and-puts-it-on-a-laptop>
> Published: 2026-11-21 03:02:29+00:00

**Key takeaway:** Inkling 2 Mini is a 26.4B-parameter open-weights model that loops its middle layers three times per token and runs locally on Macs with 32GB of RAM, scoring 20.1 on the AA Index by the lab's count.

# Thinking Machines shrinks Inkling to 26B parameters and puts it on a laptop

**Inkling 2 Mini loops its middle layers three times per token, scores 20.1 on the AA Index by the lab's count, and runs on Macs with 32GB of RAM.**

        By [Ryan Merket](https://runtimewire.com/author/ryan-merket)
        · Published 

Thinking Machines Lab released Inkling 2 Mini on November 20th, a 26.4B-parameter open-weights model built to run locally while handling agentic coding and other tool-using work. The lab, led by co-founder and CEO Mira Murati, announced the model in a thread on X and detailed it in Introducing Inkling 2 Mini.…
