It's easy to confuse counting with causing, but the tracking of model growth doesn't automatically mean OpenRouter's platform magically expanded.
This claim has a significant failure mode by implying observation directly informs or drives the expansion to 668 models from 63; that's like saying watching the water meter makes the pipes deliver more.
The actual weakest link here is ignoring the practical reasons developers might choose to integrate new models, like better APIs or reduced hosting costs, which are the true drivers.
Without those underlying incentives, a spreadsheet just shows numbers, it doesn't build anything.