This is the simplest test I could think of that tells me something about the performance of html5lib. I took three big real world HTML documents:

• the WHATWG HTML specification: http://www.whatwg.org/specs/web-apps/current-work/
• the W3C HTML Templates specification: http://www.w3.org/TR/2013/WD-html-templates-20130214/
• a diff between CPython branches: http://hg.python.org/cpython/rev/122d42d5268e:7a45415956b9

and 200 typical-size Wikipedia pages. Then I parsed them using the upcoming html5lib 1.0, a pure-python HTML parser which parses markup like your Web browser. The relevant part is this: for each of the three files I do:

etree_doc = html5lib.parse(f, treebuilder="etree")
dom_doc = html5lib.parse(f, treebuilder="dom")
assert etree_doc.tag == '{http://www.w3.org/1999/xhtml}html'
assert len(dom_doc.toxml()) > 1024

$ pip install tox
$ tox

Mac OS X 10.8.3; 2,4 GHz Intel Core i5; 8 GB RAM. Sums results from "html5lib etree" and "html5lib dom".

• Python 3.3: 874 seconds
• Python 2.7: 972 seconds
• PyPy 2.0: 195 seconds

• Python 3.3: 117 seconds
• Python 2.7: 128 seconds
• PyPy 2.0: 326 seconds