Hacker News Viewer

Inkling 2 Mini

by jasondavies on 11/20/2026, 8:12:19 PM

https://thinkingmachines.ai/news/inkling-2-mini/

Comments

by: mattnewton

Benchmarks look good, and inkling 2 small felt better than the benchmarks suggested, so I am very excited to play with this.<p>From a research perspective, I&#x27;m even more interested in the varlooped variant they&#x27;re releasing <a href="https:&#x2F;&#x2F;thinkingmachines.ai&#x2F;model-card&#x2F;tml-d24-varloop-research&#x2F;" rel="nofollow">https:&#x2F;&#x2F;thinkingmachines.ai&#x2F;model-card&#x2F;tml-d24-varloop-resea...</a> I&#x27;m not aware of any other working open source models like that.

11/20/2026, 8:21:52 PM


by: walrus01

Looking at the link to quantizations it looks like unsloth has already quantized it:<p><a href="https:&#x2F;&#x2F;huggingface.co&#x2F;thinkingmachines&#x2F;Inkling-2-Mini" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;thinkingmachines&#x2F;Inkling-2-Mini</a><p><a href="https:&#x2F;&#x2F;huggingface.co&#x2F;unsloth&#x2F;Inkling-2-Mini-GGUF" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;unsloth&#x2F;Inkling-2-Mini-GGUF</a><p>In terms of the active parameters and token&#x2F;s on slow or less capable hardware it seems like it should be in the same general performance category as qwen 3.6 35B A3B.

11/20/2026, 8:27:58 PM