The short version
Captured from openrouter.ai/pricing on 2026-11-02. OpenRouter is a pass-through: the model price is the provider’s price, and OpenRouter takes a percentage at top-up instead of per token.
| Plan | Platform fee | What you get |
|---|---|---|
| Free | None | 25+ free models, 4 free providers, chat and API access, activity logs and export, 50 requests/day limit, community support |
| Standard | 5.5% on credits purchased | 600+ models, 600+ providers, auto-routing and preferred vendor selection, budgets and spend limits, DPA via Terms of Service |
| Business | 8% on credits purchased | Everything in Standard plus managed policy enforcement, SSO/SAML, contractual SLAs |
| Enterprise | Fee discounts available | Invoicing and purchase orders through sales |
The mechanism is worth stating precisely because most readers assume the opposite: inference is billed at the provider’s list price on every plan, and the fee is charged when you buy credits - 5.5% on Standard and 8% on Business - never on individual requests. Streaming is not billed differently either; pricing is per token whether or not you stream.
The Business tier costs more for reasons that are not tokens
Paying 8% instead of 5.5% buys governance, not throughput: managed policy enforcement, SSO/SAML and contractual SLAs. On a $1,000 monthly credit purchase that gap is $25 a month, so the decision is about whether you need SSO and a contract, not about inference economics. The fee applies when credits are bought, so it is charged whether or not you spend them that month.
The free tier’s real constraint
25+ free models sounds like a free lunch; the rate limit is what bites - 50 requests per day. Enough to prototype, nowhere near enough to run a product, which is by design. Free also has no budgets, no SSO and no SLA, so it is a developer on-ramp rather than a production tier.
Payment methods differ by tier as well: Standard and Business take cards, AliPay and crypto (USDC), while Enterprise goes through invoicing and purchase orders via sales.
What the page does not say
- No per-model price table on the pricing page itself; model prices live in the model list.
- No minimum credit purchase and no stated expiry for unused credits.
- No volume discount threshold for Enterprise - only that fee discounts are available.
Reading the percentage as an effective markup
Because the fee is charged on credit purchases, the effective cost of using a model through OpenRouter is its list price multiplied by 1.055 on Standard and 1.08 on Business. On a cheap model that difference is rounded away; on an expensive one it is not. A workload spending $10,000 a month on inference pays $550 more on Standard and $800 more on Business, which is the entire justification for shopping the Business fee discount on Enterprise as soon as the spend crosses a few thousand a month.
Two practical consequences follow. First, the fee is unavoidable while the free tier’s 50 requests a day is too low for anything real, so treat 5.5% as the baseline cost of routing rather than an optional extra. Second, because the fee lands at top-up, a month with unusually high spend does not change the percentage - only the credits you buy.
FAQ
Does OpenRouter mark up model prices? No. Inference is billed at the provider’s list price on every plan; the platform fee is charged when credits are purchased.
How much is the platform fee? 5.5% on Standard and 8% on Business, charged at credit top-up rather than per request, with fee discounts available on Enterprise.
What are the Free plan limits? 25+ free models, 4 free providers and a rate limit of 50 requests per day.
Is streaming billed differently? No - pricing is per token whether or not you stream.