The increase from 63 to 668 models on OpenRouter, while a big number, doesn't automatically mean a functional daily research dataset is created.
Tracking raw growth is one thing, but unless there's a clear validation process for which models are actually useful, safe, or cost-effective, it's just data noise.
Without a defined trigger for what to do with that information, like flagging models with known security flaws or those with high inference costs, it's just a report.
For example, knowing there are 668 models doesn't help me if I still have to manually check each one for data privacy risks relevant to our members.