A flat monthly fee plus a published per-million-token rate — and never a cut of what you save. Keep your own provider keys and pay them directly, or let us manage them at cost. Either way you can model your exact bill before you send a single request.
The base fee covers the platform. Tokens are metered per million, in and out, at the tier rate — whether you route on your own provider keys or ours.
Start routing and cut your bill from day one.
For teams running AI in production.
The full optimization + governance stack.
For large orgs, security & procurement.
Turn a monthly spend into tokens, apply the levers your tier includes, and see the whole bill — provider cost at cost, plus the NeuroRoute fee. Nothing here is a share of your savings.
$149/mo base, $0.40 / 1M in and $0.70 / 1M out. Includes routing, compression and caching.
Drag to model your own workload — every figure on this page recomputes live.
Sets the assumed input : output token ratio used to turn dollars into tokens.
Provider cost cut by routing to the cheapest model that clears each task's quality bar. Every tier.
Context compression removes redundant input tokens before the request leaves us. Standard and above.
Semantic and exact-match hits are served without a provider call at all. Premium and above.
Click a tier to switch the whole calculator to it. Enterprise negotiates rates, so its saving floor matches Premium's.
Two capabilities sit outside the tier table because they are not metered per token. Neither is required to use the platform.
The chat workspace for everyone who will not call an API: 59 models behind one picker, per-person metering, and conversations in your own project. Licensed per seat on top of any tier, with a trial period — talk to us for a seat price.
Reserved throughput and self-hosted models on your own GPUs, for teams whose traffic or data rules make shared capacity the wrong answer. Scoped and priced per deployment.
