I am getting the following error with the Google Text-to-Speech API:
INVALID_ARGUMENT: `<mark>` tags are not currently supported by Studio voices.
Studio voices are GA since 26th Feb 2024 per release notes. So I don't understand what's going on here.
You are correct, but it appears that <mark> is not yet implemented for Studio voices. I would suggest filing this as Feature Request for a possibility to get implemented and gain traction here.