burnt
-synapse
ai
blog
files
status
speedtest
mail
← all notes
#mtp
every note tagged mtp
date
note
tags
22 Aug 2026
mtp: 60 t/s when the text is predictable
the box does 35 t/s of generation. with mtp — multi-token prediction — it does 60+ when the next few tokens are easy to guess.
#qwen
#llama.cpp
#mtp