> and Clef-flash is based on Qwen3.8-9B

There is no official qwen 3.8 9b

From the model card:

> Clef-Flash is post-trained from Qwen/Qwen3.5-9B. See Clef for the larger variant.

Thanks, I missed that. Fixed my comment.

It's based on Qwen3.5-9B, maybe a typo

I've been guilty of the same wishful projection

Atom is 60M Param (around 133x to 400x smaller).

16ms latency. And locally run.

https://at0m.pienomial.com/

Why go big when you can go small ?

Because with a tiny model you're skipping all the intelligence and world knowledge that makes it useful without fine tuning. `typed-decisions` is almost entirely text classification tasks.

Don't assume. It has generic world knowledge. It is performing well on the benchmarks we didn't even train it on.

Cause it's not open?

Good point.

To counter, most of the AI is not open. So is none of Microsoft Products. As long as they work, we keep using them.

counter point, I've stopped using all closed models and harnesses as a life choice

Ai is too important and transformational to let Big Ai dominate in a closed ecosystem, thankfully the Chinese have a different mindset and approach

That's an absolutely fair way. More power to you !