I entered HTML5 elements – specifically figure and figcaption in the HTML mode of the tinymce editor. Save the post and view on the front-end and it works – the HTML5 elements are rendered properly.

However, if I go back to the tinymce editor and switch to the Visual mode, it breaks the HTML5 markup. In this particular case, the figure elements becomes empty -> <figure></figure> — and all the markup that was previously saved inside figure is now outside of it and immediately after it in the source as well as the HTML mode if I go back there.

Seems to be a conflict with how the Visual mode parses HTML5. Any ideas how to fix this?