Except for being free, WebM was not a "superior compression format" to H264 last I heard. Has it changed recently?

I think it all but caught up lately in VP8 vs h.264 performance. Google kept improving VP8 performance. However, webm has also added VP9 support recently, which is clearly superior.

This is not true. My work, written three years ago, showed that VP8 was inferior to H.264[1] at the time. A more recent paper showed that both VP8 and VP9 are inferior to both H.264 and HEVC[2].

[1]: http://blog.seancassidy.me/h264-and-vp8-compared.html

[2]: http://iphome.hhi.de/marpe/download/Performance_HEVC_VP9_X26...

Is VP9 in much use (compared to VP8)? I've definitely heard good things about it, but I have no idea how well supported it is or whether anybody has it in use yet.

Superior in what sense? Not quality or file size.

Well, it's patents are free to licence and with open source code, so it's superior in a way that for many (me included) believe trump the others.

Webm is superior to gif in terms of quality AND file size. (Other video formats, not so much).

Superior to GIF.

Pedantry: WebM is not a compression format. It is a container format. At the moment it appears to support VP8 video, VP9 video, Vorbis audio, and Opus audio; there is no single form of compression that WebM implies.