Skip to content
Better HN
Top
New
Best
Ask
Show
Jobs
Search
⌘K
Experimental 3.5T K2 merge beats GPT-4.5 and Opus at writing | Better HN
Experimental 3.5T K2 merge beats GPT-4.5 and Opus at writing
(opens in new tab)
(huggingface.co)
2 points
bo0tzz
5mo ago
1 comments
Share
1 comments
default
newest
oldest
mynti
5mo ago
"beats at writing" is kind of far fetched since the page provides three generic pormpts with answers from Opus, GPT5 and K3. But a 3.5T model is crazy huge and the answers still feel more or less the same to the other models
j
/
k
navigate · click thread line to collapse
