Benchmarks
Daily measurements from 6 locations, organized by category.
Real, paid calls to AI model APIs, direct and through OpenRouter, timed to the first visible token.
APIs that read a piece of text and return a structured answer to a question, rather than streaming prose. Each gets the same yes/no question, and we time the whole JSON response.
Show in the table →Public endpoints of APIs an application calls at runtime.
Show in the table →Package registries, code hosting and the websites developers work in.
Platforms that host and deliver websites, measured by pages they serve themselves.
Show in the table →Stylesheets, scripts and other assets a browser fetches from a shared CDN.
Show in the table →Big public sites used as a delivery reference, not as alternatives to each other.
Show in the table →Software and commerce platforms whose product sits behind a login, measured by the site they serve.
Show in the table →Free public JSON APIs that tutorials and prototypes call.
Show in the table →35 benchmarks shown
* Time to first token. Everything else: total response time.
** Fastest to slowest location per day.
Listed alphabetically.
Run a free test from the same 6 cities and see where the time goes.
Free, no account needed.