How billing works

FluxRouter is pay for what you route. You are billed for the tokens you actually send and receive, at the rate of the model alias you send. There is no per-seat fee and no minimum commitment to send a request.

Current rates for every lane, every pinned model and every ability are on the pricing page. This page explains how they apply; it does not repeat the numbers, so it never goes out of date.

How is a request priced?#

Every model request is metered by tokens. Input tokens (your prompt, including history and tool definitions) and output tokens (the reply) are priced separately, and output costs more than input.

What rate applies depends on the model you send:

The Pay As You Go plan adds a small markup to the lane rates; the pricing page shows the rates with it applied. Builder (and Scale, for accounts already on it) bills usage above the included credit at the plain lane rates.

How does flux-auto bill?#

flux-auto picks among vetted models for each kind of task, cost-aware. It bills at the Standard rate, or the Fast rate when it sends your request to the Fast lane and a Fast model answers. It never bills more than Standard: a hard prompt that reaches the Reasoning lane still bills at the Standard rate. If a request starts on the Fast lane but a model from another lane gives the answer, it bills at the Standard rate. If you know a job is simple and high-volume, send flux-fast to pay the Fast rate on every request. See Let Flux pick, or pin a model.

What do abilities cost?#

Ability calls bill per use, not per token: per page read, per image, per second of audio transcribed. Each is billed exactly as the direct API call would be, whether your agent makes it through the API or through the Flux tools server. Prices are on the pricing page.

How do I see what a request cost?#

Every response tells you which model answered. The cost of a non-streaming request is in the X-Flux-Cost-Usd response header; on a streamed request it arrives in the final usage block instead. Your running total is at /home/usage. See Receipts and Usage and invoices.

Are there seats or minimums?#

No. There is no per-seat fee and no minimum spend. That holds on the Free plan (within its cap), on Pay As You Go, and for usage above a paid plan's included credit.

What about plans with an included credit?#

Builder (and Scale, for accounts already on it) includes an amount of routing each month. The included credit refills on the 1st of each month (UTC), not on the day your plan renews: the renewal date is when the plan fee is invoiced. Usage up to the included credit is covered by the plan fee; usage above it bills at the lane rates. The metering is the same either way. See Plans.

How do I keep costs down?#

• Send flux-fast for simple, high-volume work.
• Keep prompts and history lean: input tokens count on every turn.
• Cap max_tokens where long replies are not needed.
• Set a spend cap so a runaway loop stops early. See Spend caps.