Billing

FluxRouter meters the tokens you route, at the rate of the alias you send, and ability calls (page reading, images, transcription) at their own per-use rates. There are no seats and no minimum to send a request. Pay as you go is the plan for new accounts: you top up credit, every call is charged at cost plus 12%, credit never expires, and calls stop when your credit runs out. Current rates are always on the pricing page, which reads them from the same configuration the gateway bills from.

What's in this section#

• How billing works: what a request costs, how flux-auto and pinned models bill, and how to keep costs down.
• Plans: Free, Pay As You Go and Enterprise, and what each one adds.
• The Free plan: the credit and default key you get at sign-up, what it covers, and what happens at the cap.
• Spend caps: the limits that stop a request, and the ones you can set yourself.
• Usage and invoices: where to see spend, the per-request cost, and your invoices.

If a request came back with HTTP 402, the error reference lists every refusal code and its fix.

The short version#

• You pay per token at the rate of the alias you send. Input and output are priced separately.
• flux-auto 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. For the lowest rate on every request, send flux-fast.
• A pinned flux-pinned-* model bills at that model's provider rate plus the Flux markup.
• Every new account starts on the Free plan with a small monthly credit and a key made for you, and stops at the cap rather than billing you.
• Your key never changes when you change plan.
• One key gives you everything Flux does, at the same pay-as-you-go rates: every model, smart routing, image generation, voice, web search and tools. There is no plan to upgrade to.