Platform features
Platform features
Routing, cache protection, and data handling that the gateway adds on top of the model APIs.
Modelflare is more than a pass-through proxy. Five platform behaviours decide how a request is routed, what it costs when a route fails, what happens when prompt caching underperforms, how long the record of it is kept, and how a team shares one balance. They apply to every endpoint documented on this site and need no change to your client code.
Smart API Key
One key for every model your account can use, with a routing strategy instead of a fixed group.
Group failover
A primary group plus ordered fallback groups, so one unavailable route does not fail the request.
Cache hit guarantee
Protected OpenAI groups carry a minimum daily cache hit rate; a shortfall is credited back.
Data retention
How long logs and failure diagnostics are kept, and how to shorten it for your account.
Enterprise accounts
One shared balance for a team, with prefixed keys and a spending budget per member.
Prices, availability, and the group each model is enabled on stay on Models & Pricing.