Every figure on this site comes from somewhere. Here is where, how often it refreshes, and what we do when a provider changes the rules mid-month.
Taken from each provider's published price list and converted into credits at 1,000 credits per dollar. When a provider prices per million tokens, we hold that unit rather than rounding to a per-message figure, and the per-message numbers shown elsewhere assume 1,000 tokens in and 500 out.
Funding rates are quoted at signature time from a median across major venues. Nothing on the pricing page is a promise about the rate you will get later; it is the rate as of the moment the page loaded.
Standard individual tiers, billed monthly. Annual discounts, team seats and regional pricing are ignored deliberately, because including them would flatter the comparison in our favour.
"341 models" is the size of the served catalogue. "185 run in the workspace" is the subset wired to the chat surface; the rest are callable by API or listed for reference. Both move without notice as providers add and retire models.
If a number here is wrong, it is worth telling us, contact. And it gets corrected rather than quietly adjusted.