Model catalog

Compare models across every provider we route to: context length, price per million tokens, quantization and the endpoints actually serving each one.