Model Support for kimi-k2-turbo-preview via the Moonshot AI API #7078
khaosoi
started this conversation in
Feature Requests
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
WHAT?
This is a higher speed version of kimi-k2 (40 tps vs 10 tps for regular kimi-k2-0711-preview). Appears to be available only via the Moonshot AI API. Cannot find it listed on OpenRouter.
WHY?
One of the biggest complaints about kimi-k2 is that it is slow. Otherwise this is a good model for RooCode and it would be good to have an uncompromised (*) faster version of it available to experiment with.
*Yes there fast offerings of regular kimi-k2-0711-preview available on OpenRouter, but they come with aggressive quantizations -- which we don't really want, do we?
Beta Was this translation helpful? Give feedback.
All reactions