Hacker News
new
past
comments
ask
show
jobs
points
by
andai
18 hours ago
|
comments
by
dannyw
15 hours ago
|
next
[-]
It’s a 6bn model. Totally different class. I’m more excited about “frontier small language models” tbh.
reply
by
andai
5 hours ago
|
parent
|
[-]
It's a 119B model, 6B active.
That's still 3-10x smaller than the other models in that graph though (400B, 1T, 1.5T).
reply
by
rtaylorgarlock
17 hours ago
|
prev
|
[-]
Agreed, though open weights + relatively small is still headline worthy. This thing really cooks.
reply