From what I've read online it's not necessarily a unquantized version, it seems to go through longer reasoning traces and runs multiple reasoning traces at once. Probably overkill for most tasks.

Yup, that was it. Didn't realize they're different models. I suppose naming has never been OpenAI's strong suit.

>It's the one you have access to with the top ~$200 subscription and it's available through the API for a MUCH higher price ($2.5/$15 vs $30/$180 for 5.4 per 1M tokens), but the performance improvement is marginal.

The performance improvement isn't marginal if you're doing something particularly novel/difficult.