Exactly. Without Claude and GPT, we'd be limited to Grok. A lot of ink has been spilled telling us exactly how to use Sonnet, Opus, Sol or Luna in certain scenarios, Grok hasn't ever been recommended by our pointy haired friends even once.
Well isn't that going to be an interesting test of corporate ideology ;-)
It will be a test of what they think about the relative cost / performance of a model is, and not a single thing more than that.