HN
Hacker News
Top
New
Best
Ask
Show
Jobs
Comment by dalant979 | Hacker News Reader
Parent
Full thread
dalant979
·
180B?
View on HN
kzrdude
·
125B regular params, 51B engrams, 4B MTP. Something like that. It should have a label of effectively 125B params with A6B (6B active).
Reply on news.ycombinator.com