5.06% here (Austria) by IP addresses (10440 out of 206384 in 24h).

That's not counting those Opera users who switched their User-Agent string to MSIE or similar because of stupid JS detection like MySpace's.

In my experience, Opera gets significant usage in non-English-speaking countries. I read at one point (several years ago, so grain of salt) that 25% of Russian users were using Opera.

That said, our own traffic shows 0.80% Opera users, which puts it just above RockMelt. It's silly to actively filter them out as MySpace is doing here, but would be equally as silly (for us at least) to devote any resources to dealing with any Opera-specific issues, until the usage stats justify it.

Yes, if you look at http://gs.statcounter.com/ and select individual countries you can see that Opera is strong in Eastern Europe and parts of Asia. There are historical reasons for that (just my opinion, see bellow) and also the fact that Opera had/has Turbo - which really helps on slow connections (mobile).

So if you run US-focused site, there isn't really reason to care about Opera, and test for it. Blocking it (or any other browser) is bad though.

The "historical reasons": Opera was by any means the best (serious) browser out there, but it was paid. People in these countries weren't paying for software anyway, because their income was well bellow western standards while prices were same, and as a bonus, there was no one who would care. So while someone on the west would rather use something free (like IE), people in (for example) Ukraine picked the best. That's why first decent browser (Firefox) took over the west so fast.

My site shows Opera at 4%, above Safari at 3.6%.

I use Opera Mini (or is it mobile?) on my old phone - they have a system that routes request through their own caches that downscales/downsamples images, removes cruft, etc. before the pages hit the mobile device. I imagine that means they get under-represented in hit figures (but not by that much).

Looking at hits is doing it wrong - profit is where it's at, what proportion of profits come from Opera users.

[For example, hypothetically, they might be highly paid trend setters and amount for a disproportionately large profit share on a fashion site or such.]

I'm showing Opera as 1.61%. Yeah, still not exactly massive.