Dataset 20261011-0550-f160c57 · measured 2026-10-11 05:50 UTC

MathML — presentation markup

188 tests judged in this suite · 1 measurement tie, counted as a failure.

96.3%181 of 188 subtests

By tests — the unit Interop counts — 96.3%: 181 of 188. A file is one test and passes only when every subtest in it passes, so this figure is the stricter of the two.

181Passingrendered pixel-identical to the reference, within the test's own fuzz allowance

6Failingevery one of them listed below, with the evidence

0Not judgedneeds JavaScript, interaction or a draft feature — a PDF engine has none

Every test in this suite

All 188 of them, grouped by the directory the corpus keeps them in — passes included. Open a directory to see what each test checks; open a test to see our render, the reference, the pixels that differ, and the source of both files.

188 tests in 11 directories

These are web-platform-tests files, unchanged. How the browsers score on the same suite · the suite upstream

Dataset 20261011-0550-f160c57 · measured 2026-10-11 05:50 UTC

Loading…