Tiers and pricing lanes
The three FluxRouter pricing lanes and what the flux-fast, flux-standard, and flux-reasoning tier aliases mean as user-facing choices.
FluxRouter groups models into three pricing lanes, and gives you three tier aliases that map onto them. This page explains both, and how they relate.
What are the three pricing lanes?
Every model on the gateway is priced in one of three lanes. You pay per token at the lane rate of whichever model actually served your request. All rates are pay-as-you-go, with no minimum commitment.
| Lane | Starts at | Good for |
|---|---|---|
| Express Lane | $1 / 1M tokens | Lightweight, high-volume, latency-sensitive work |
| Daily Driver | $2 / 1M tokens | General coding, writing, and analysis |
| Deep Thought | $4 / 1M tokens | The hardest reasoning and frontier-model work |
"Starts at" is the floor for that lane. Individual models within a lane bill at their own published rate, and the rate you are charged is the live rate for the model that served the request.
What do the tier aliases mean?
A tier alias is a way to ask for a class of model without naming a specific one. Instead of sending flux-auto and letting Flux decide, you send the tier you want:
flux-fast— the fast, inexpensive tier. Use it for simple, high-volume, or latency-sensitive requests where a lightweight model is enough.flux-standard— the general-purpose tier. A solid balance of capability and cost for everyday coding, writing, and analysis.flux-reasoning— the strongest tier. Use it for hard problems, complex reasoning, and tasks where you want the most capable model.
There is also flux-image for image generation, and flux-pinned-* aliases for locking in one exact model. The authoritative list of everything you can send is GET /v1/models.
How do tiers and lanes relate?
The tier aliases line up with the pricing lanes: a lighter tier is served by a model in a cheaper lane, and a stronger tier by a model in a more expensive lane. So choosing a tier is also choosing roughly what you pay. You always see the actual model and its cost on every response through the transparency headers.
Which should I use?
- Want good results without thinking about it? Send
flux-autoand let Flux route. See How smart routing works. - Want to steer toward cheaper or stronger without naming a model? Send a tier alias.
- Want the exact same model every time? Pin a
flux-pinned-*id. See flux-auto vs pinning a model.