The jump from 63 to 668 models on OpenRouter sounds impressive, but pure growth numbers don't automatically mean meaningful insight; it's a scarce resource if we're just talking inventory.
Tracking how many models exist doesn't tell us about their actual utility or performance, which is what truly matters when resources are dwindling.
It's like having a huge applicant pool: you still need to screen for quality candidates, not just count the résumés.
Our attention and budget are finite, so knowing the sheer quantity of models doesn't inform strategy unless we know which ones actually deliver. For example, if 600 of those models are niche or low-performing, then the initial count is just noise, not data for actionable decisions.