logoalt Hacker News

cpburns2009yesterday at 4:00 PM2 repliesview on HN

This is the best I got, all with Unsloth's quantizations.

Laguna-S-2.1:UD-Q4_K_XL (no MTP) pp=186.4 t/s tg=27.8 t/s

Qwen3.6-35B:UD-Q4_K_XL (with MTP) pp=404.4 t/s tg=83.2 t/s

Qwen3.6-27B:UD-Q4_K_XL (recorded pre-MTP) pp=343 t/s tg=12.1 t/s

Laguna actually performed better than I remembered. I thought it was slower.


Replies

htrptoday at 1:51 AM

Have you benchmarked against full precision models for accuracy/ performance?

show 1 reply