It's more like they released GPT 6 Sol too early because they were under pressure and now they are releasing the real version. You cannot do anything more than minor post-training in a week.

Implying they don't have like 3 or 4 "models" (different quants, post training, plain renaming) on the back burner at any point in time to do exactly that

I think this was supposed to be new Astra, but it came out shit. So instead of shelving it, they found a way to make it still seem like progress.

It's a competitive strategy with Anthropic, not necessarily releasing as soon as they're done

>That seems like...a big deal.

They can release a new version every day if they wanted to. The question is whether or not the new releases provide substantial improvements or not. It's not hard to just go through the motions, bump the minor version, then make an announcement to rile up the users who don't get that none of this is standardized or regulated in any way and it's literally all made up by the company trying to sell them the product.