I mean we do know the formula, se we know pretty well how accurate it was, since we can just use the same formula with meters and calculate it ourselves

"The simplified method works by considering two cities along the same meridian and measuring both the distance between them and the difference in angles of the shadows cast by the sun on a vertical rod (a gnomon) in each city at noon on the summer solstice. The two cities used were Alexandria and Syene (modern Aswan), and the distance between the cities was measured by professional bematists.[16] A geometric calculation reveals that the circumference of the Earth is the distance between the two cities divided by the difference in shadow angles expressed as a fraction of one turn. "

Sorta. The stade was based on the pous which varied. But not that much, and in ways that are often consistently documented. Around the time Eratosthenes was alive, give or take a couple hundred years, it was documented that 1 Roman mile = 8 stades, which gives us something to go of off. While there are other possible definitions, we do have one that we know is probably closest to whatever Eratosthenes used.

EDIT - the numbers regarding the error range in this source is likely inaccurate, but goes into the units issue

https://maa.org/press/periodicals/convergence/eratosthenes-and-the-mystery-of-the-stades-how-long-is-a-stade