On: Run Qwen 3.8 Flash Next (125B) on consumer hardware (RTX 4090) at 100T/s

[snehesht]

Yeah I agree, I'm running it with Pi didn't notice much difference compared to lower tier models and the speed, of course.

[nicce → snehesht]

I am running 27B with Deepseek Harness these days and somehow just by using it, without any parameter changes, the model feels even more intelligent.

[DoctorOetker → nicce]

do LLMs tend to be homesick when not used in the same harness they sat in during some training phase?

[Bnjoroge → DoctorOetker]

iirc there was a sectionin Qwen’s paper where they talked anout how they post-trained flash or 3.8 to work just as well regardless of the harness or eval used. I think that used to be true but not sure if it is any longer