Rankings

Models ordered by things that can be measured from what the gateways publish: context window, price, free availability, declared parameters, and how far gateways disagree on the same model.

There is no "best model" page here, and there will not be one. Every ranking on this site sorts a column that exists in the gateway catalogs. We hold no benchmark scores, no quality evaluations, and no licence data, so a "best for coding" or "best open-source" list would be an opinion presented as a measurement. How this data is collected.

LLMs with the largest context windows

1,040 models publish a context window across the gateways tracked here, and 454 of them accept a million tokens or more. The largest is Auto Router…

1,040 models Ranked on context window

Cheapest LLM APIs by input price

1,043 models have a published input price above zero. The cheapest is Gemini 3.5 Transcribe Live at $0.0001 per million input tokens. Prices are the…

1,043 models Ranked on input price

LLMs available free through a gateway

200 models are served at zero cost by at least one tracked gateway. Free here means a gateway publishes a price of zero — not that it is unmetered…

200 models Ranked on free availability

LLMs that support tool calling

526 models are declared by their gateway as accepting tool-calling or structured-output parameters — the prerequisite for using one in an agent loop.

526 models Ranked on declared parameters

The same model at different prices

210 models are priced by more than one gateway, and they do not agree. The widest gap is Mistral Nemo: 8.95x between its cheapest and dearest gateway…

210 models Ranked on price spread