METHODOLOGY
Where our data comes from
A short note on how we put AIModelFind together, so you know how far to trust what you see.
Prices
Prices come from the model makers and the providers themselves: their price pages, APIs and published price lists. Third-party lists only help us discover models; we don’t show their prices.
Each provider bills in its own units, so we show prices the way they charge, per 1M tokens, per image, per second and so on. When a price depends on settings like resolution or context length, we show the standard rate or the lowest tier, marked from.
How often
We refresh the data regularly, and each model page shows when its prices were last checked.
Keeping it accurate
Before anything is published we run automatic checks that catch broken sources and numbers that don’t add up, and suspicious prices are reviewed rather than shown.
When two official sources disagree, we show both and say where each comes from.
Specs, open weights and scores
Specs such as context length and release dates come from the makers first. Open weights and licenses come from the makers’ own model pages. Quality scores come from public leaderboards such as LMArena, credited where they appear.
What we leave out
We leave out prices we can’t confirm or whose unit is unclear, and sources that don’t allow their data to be used.
Independence
We don’t sell model access. Any sponsored or affiliate links are labelled.
Found a mistake?
Use the report link beside any price or spec, or the contact page, and we’ll look into it. Contact