Laguna M.1 is the flagship coding agent model from Poolside, optimized for complex software engineering tasks. $0 per million input tokens, $0 per million output tokens. 131,072 token context window, maximum output of 8,192 tokens.
#Poolside Laguna M.1
27 messages · Page 1 of 1 (latest)
that's a low context window for a coding model
8k response is yikes
seen single tool calls bigger than that
eager to try it though, new coding models always exciting
btw their official api is free, so it might be less ratelimited
aww their cli isnt open source
https://github.com/poolsideai/pool
Hahaha La Goona
I'd hope the tiny context/output size is only because it's the free version
and maybe they have a Laguna L1 which is huge?
The small one, XS.2 seems like the good one
well I've seen worse one-shots
after a while it devolves into explosively-reproducing predators who eat everything including each other but I guess that's a consequence of the rules
the snail trail is so unnecessary but kinda neat, first time I've seen that unprompted
I'd say it's kinda dumb/hopeless when it comes to inventing UIs on its own
can't even make a functional abacus smh
The model page is now showing higher context and output lengths than it did initially, looks like
This is what was announced when it was released
went from useless to useless 🥳