Re:Inkling 2 Mini
by hashmap on Saturday November 21, 2026 @12:32AM
(#50586262)
yeap, and 3x is where i saw the most gain from it too. you can set up most any model to do just this if you can find its reasoning layer boundaries, like qwen3.8 for example, h27-h35 will do it.
[ Parent
| on HN ]
There are no comments on this comment yet.