Model Hub
The right model for each job.
The Model Hub offers 30+ hosted and open-weight LLMs in one catalog, plus routing and settings for access, safety and cost.
3 trillion+
Monthly tokens handled by Acme
64%
Less token spend than a one-model setup
30+
Hosted and open-weight models on offer
A single catalog for picking and managing the models you use.
Switch providers freely
Use leading hosted and open-weight models side by side, and adopt new releases without renegotiating anything.
Decide who uses what
Turn individual models on or off, choose who may call them, and cap usage so costs stay in line as more teams join.
Same protection on every model
Injection checks and runtime data safeguards apply the same way whether a request goes to a hosted model or an open-weight one.
Auto routing
BetaLet Acme pick the model for each task.
Routing scores 30+ LLMs on quality and price for every request, so you get top-tier answers on hard tasks and cheaper models on the easy ones.
See auto routingGovernance
Your catalog, your rules.
Usage controls let admins enable or retire models, grant access by person or team, and set spending caps for each person, agent, or department.
See usage controlsAI Gateway
One policy across every AI app.
The AI Gateway applies the hub's access rules and budgets to model traffic from other AI apps too, not just requests that start in Acme.
See the AI GatewayAcme Shared Model Key
BetaA single key for many providers.
Call top hosted and open-weight models from several vendors with no data retained. Nothing you send trains a model, and the same safeguards apply to all of them.