Here are Cursor’s default enabled models.
May you explain what is the difference between cursor-fast and cursor-small?
Likewise, what is the difference between o1 and o1-preview (which was supposed to be just a temporary version until the full release of o1)?
o1 is the new version from Open AI, its the successor of preview
though to be fair its even more censored and limited than 01-preview and seems to have lost some intelligence in quite a few areas because of that
though for coding o1 mini is still better than both
usually o1 is good for reasoning and creative stuff, but yet again, the preview is more creative than the fully released version
about cursor small and Fast i bet some DEVs will answer soon
would be nice if they had a decent blogpost about recent changes, the changlog on the web is quite limited
cursor-small and cursor-fast were two internal models we created ourselves to trial out! However, we are replacing cursor-small with cursor-fast soon, but keeping the name cursor-small, so there will only be the one option soon!
o1-preview was the demo version of o1, and may perform slightly worse. Both models have the same usage cost, so I’d preference o1 over the preview model unless you see a big difference.
Thanks! So if you remove the selectable option of cursor-fast soon, will you likewise remove o1-preview and keep just o1?
I’d imagine so, unless we see any difference between the models in the future.
cursor-small based on LLama3-8b or?
It’s our own in house model, but I won’t reveal any secret sauce for now! We may have a blog post about it in the future with more details, but that’s all I can say for now… 
I feel cursor-fast is faster than cursor-small.
