Pricing
One rate. Every model.
Pay exactly what the model provider charges, plus a flat 1% platform fee. No tiers, no per-model markup, no minimum spend.
Platform fee
1%
Applied to the provider's list price on every request, for every model — the one you asked for, or the one model="auto" routed you to.
OpenRouter's ~5.5% is a deposit fee, charged when you add credits. Our 1% is an inference fee, charged only on what you use.
No tiers
Every account, every model, the same 1% — whether you send ten requests a day or ten million.
No markup
You're billed at the provider's list price plus the 1% fee — never a hidden spread on top of it.
No minimums
No monthly plans, no seat fees, no commitments. Pay for what you use, billed per request.
Worked examples
What 1% looks like in practice.
Prices below are per 1M output tokens at each provider's list rate. LLMRouter passes through the provider cost and adds 1%.
| Model | Provider price | +1% fee | You pay |
|---|---|---|---|
| Llama 3 8B | $0.20 | $0.002 | $0.202 |
| Mistral Small | $3.00 | $0.03 | $3.03 |
| Claude Haiku | $4.00 | $0.04 | $4.04 |
| GPT-4o | $10.00 | $0.10 | $10.10 |
| Claude Opus | $75.00 | $0.75 | $75.75 |
Illustrative provider rates, per 1M output tokens — check the dashboard for live, per-model pricing.
With model="auto", most of your traffic routes to the cheap end of this table — the 1% fee only ever applies to whatever model actually handled the request, so smart routing compounds with the low fee instead of fighting it.
Vs. the alternative
A cheaper fee, and it's not the same kind of fee.
OpenRouter
~5.5%
deposit / Stripe fee, charged when you add credits — inference itself is billed at 0% markup.
LLMRouter
1%
inference fee, charged only on what you actually use — itemized on every request.
These aren't the same charge, so they don't compare 1:1 — but on $10,000/mo of model spend, a 1% inference take-rate is $100, well under what OpenRouter's deposit fee costs to fund the same spend. And that's before smart routing starts cutting your underlying provider bill by 40–70%. We also charge a separate, disclosed Stripe fee on deposits — priced at cost to cover card processing, never absorbed into the 1%.
FAQ
Pricing questions, answered.
Is the 1% fee the same for every model?
Yes. Whether a request lands on a small open model or a frontier reasoning model — and whether you picked the model yourself or let model="auto" route it — the fee is always 1% of the provider's list price for that request. No per-model exceptions.
How is the 1% calculated and billed?
Each response is billed at cost — the exact input and output token price the underlying provider charges — plus 1% of that amount. It's itemized per request in your traces, so you can see the provider cost and the fee separately.
Do I need my own provider API keys?
No. LLMRouter holds provider access on your behalf, so you get a single API key and a single bill. If you'd rather bring your own provider keys, that's supported too — the 1% fee still applies to routed usage.
Are there monthly plans or seat fees?
No. There's no subscription, no minimum spend, and no per-seat pricing. You're billed only for what you use: provider cost plus 1%.
Start routing in minutes.
Create a developer account, grab an API key, and point your OpenAI SDK at us.
OpenAI-compatible · one API key · no code rewrite.