API free tiers and rate limits

A script reads each provider's own rate limit page. Tables we can read directly are parsed in code; the rest are read by a language model told to copy each value exactly as written. Every value is then checked against the page text, and any value, tier or model name the page does not print is held for review instead of published. A free tier is marked yes only while the page itself says one exists; "not stated" means the page does not say, not that there is none.

Some providers publish per-model limits only inside their console (Google Gemini, Mistral) or set them per account; those tables show what the public page prints, such as tier spend caps. Limits change without notice, and your own account may differ: check your console before you plan around a number. A new row is recorded only when a value is first seen or changes, with the page address and the time it was read.

Prices for the same models are in the LLM API pricing timeline. More on how we source figures: how we collect our data. Spotted a wrong value? Write to ops@3ammarketer.com.