It does stay responsive while scanning a large (200k) [Gmail]/All mail folder, but apparently it has a single connection to Gmail because no other folders or messages can be obtained during the download. After about five minutes it has a basic list of headers (but not the subjects) and there seems to be more concurrency. It would be better if it chunked the large scan and allowed other operations to complete (assuming progressive scanning is possible).
Interesting, because the only slow action when opening mailboxes is fetching and applying the thread index, and gmail does not support that particular feature. Which means I don't really know how it can take a "few minutes" on GMail -- all it does it simply fetching the UIDs...
I wanted to know the same, so I looked it up on Wikipedia:
Formerly known as Sylpheed-Claws, it [Claws] started in April 2001 as the development version of Sylpheed, where new features could be tested and debugged, but evolved enough to now be a completely separate program. It forked completely from Sylpheed in August 2005.