Lanes
The three Flux text lanes, Fast, Standard and Reasoning, and what the flux-fast, flux-standard and flux-reasoning aliases mean as choices.
Flux groups its text models into three lanes, and gives you one alias for each. This page explains both, and how they relate to what you pay.
What are the three lanes?
Every text model on the gateway sits in one of three lanes. Each lane has its own per-token rate, with input and output priced separately. Current rates are on Pricing, which reads them from the same pricing data the gateway bills with.
| Lane | Alias | Good for |
|---|---|---|
| Fast | flux-fast | Lightweight, high-volume and latency-sensitive work |
| Standard | flux-standard | General coding, writing and analysis |
| Reasoning | flux-reasoning | The hardest reasoning and frontier-model work |
What do the lane aliases mean?
A lane alias asks for a class of model without naming one. Instead of sending flux-auto and letting Flux decide, you send the lane you want:
flux-fast: the fast, inexpensive lane. Use it for simple, high-volume or latency-sensitive requests where a lightweight model is enough.flux-standard: the general-purpose lane. A sound balance of capability and cost for everyday coding, writing and analysis.flux-reasoning: the strongest lane. Use it for hard problems, complex reasoning and tasks where you want the most capable model.
A lane alias bills at that lane's rate. Choosing a lane is also choosing what you pay.
How do flux-auto and pinned models bill?
flux-autopicks a lane and a model for you. It bills at the Standard rate, or the Fast rate when it sends your request to the Fast lane and a Fast model answers. Never more than Standard.flux-pinned-*ids lock in one exact model and bill at that model's provider rate plus the Flux margin shown on Pricing.- On the pay-as-you-go plan, a markup is added to these rates. How billing works explains it.
You see the actual model and the cost of each request in the receipt headers.
Is images a lane?
No. Image generation is a separate ability with its own endpoint and per-image prices, not a text lane. See Make and edit images.
Which should I use?
- Want sound results without thinking about it? Send
flux-auto. See How routing works. - Want to steer towards cheaper or stronger without naming a model? Send a lane alias.
- Want the same model every time? Pin a
flux-pinned-*id. See Let Flux pick, or pin a model.
The authoritative list of everything you can send is GET https://api.fluxrouter.ai/v1/models, and Models shows it in the browser.