Gemma-4-26B-A4B with 39-69% less thinking at same accuracy

Re:Gemma-4-26B-A4B with 39-69% less thinking at same accuracy by Foobar8568 on Friday October 23, 2026 @02:51PM (#50213678)

But Qwen 3.8 27b is unbearable slow, you get 10minutes of "let me think" over any trivial stuff.

[ Parent | on HN ]

Re:Gemma-4-26B-A4B with 39-69% less thinking at same accuracy by simonw on Friday October 23, 2026 @02:55PM (#50213733)

Qwen 3.8 27B defaults to "xhigh" (its maximum level) for thinking, which makes it diabolically slow. Setting it to "medium" or "low" speeds it up a lot.

[ Parent | on HN ]

Re:Gemma-4-26B-A4B with 39-69% less thinking at same accuracy by dofm on Friday October 23, 2026 @03:12PM (#50213946)

It does. It's quite good with the low setting. I prefer Muse Glimmer for, I dunno, hard to explain comfort reasons.

I had to hack the template of Qwen 3.8-27B to get it to go into low reasoning mode. Not sure if that is fixed.

Have you tried Nemotron 74B-9B? (Slow on Mac with llama.cpp due to a missing tensor type)

[ Parent | on HN ]

Re:Gemma-4-26B-A4B with 39-69% less thinking at same accuracy by electroglyph on Friday October 23, 2026 @04:33PM (#50214939)

that nemotron model is such a dog, like all the other nemotrons

[ Parent | on HN ]

Re:Gemma-4-26B-A4B with 39-69% less thinking at same accuracy by dofm on Friday October 23, 2026 @05:01PM (#50215270)

I have found it to be OK on sort of general purpose overview code stuff. At the moment it's too slow in llama.cpp on Mac for me to bother testing further.

[ Parent | on HN ]