HomePricing
Pricing
Nothing is billed upfront. Each model meters in its own unit — tokens, images, seconds or characters — and the provider list price is printed beside ours on every row.
Wallet Balance
¥0.00
Text models bill input and output tokens separately. Image models bill per image, video per second, audio per 1,000 characters or per track.
Bonus Balance
¥0.00
LrSite.p__how__2d
How the meter works
There is no plan to pick and no per-seat charge. You buy credit, it meters down as you call, and whatever is left stays yours.
/01
Per-unit, not per-seat
LrSite.p__how__1d
/02
Credits never expire
LrSite.p__how__2d
/03
You see the margin
LrSite.p__how__3d
Recharge packages
The more you load, the lower your effective rate. Bonus credit is added on top of the amount you pay.
Card top-ups open soon. These are the tiers you’ll be able to pick.
Custom amount
LrSite.p__custom__d
Bonus tier+10%
Credit received (est.)≈ ¥165.00
LeapRiver vs going direct
An honest comparison. Direct access wins on a couple of rows — we would rather you know that upfront.
| Feature | LeapRiver | Direct provider | Other relays |
|---|---|---|---|
| Price vs provider list | 8–12% under | List price | ~5% over |
| Both prices shown upfront | Yes | Yes | Often hidden |
| Automatic cross-provider failover | Yes | Build it yourself | Varies |
| Single key & invoice for 30+ models | Yes | No | Yes |
| Bring your own provider keys | Business plan | n/a | Varies |
| Provider-specific features (batch, caching) | Most | All | Some |
Pricing questions
Questions we get asked
The short answers, before you open a support ticket.
Where does the provider list price next to each model come from?
From the public pricing pages of the providers we route to, refreshed every 6 hours. The number is what the provider charges on a US-dollar, US-East, pay-as-you-go account at the moment of the refresh. We display it so you can see the saving without opening another tab.
Is the discount the same for every model?
No — it varies by model because the routing mix varies. Newer models often have a wider gap because the list price moves faster than the blended cost. You can pin a specific model on the models page and watch its rate refresh.
Is batch cheaper than online?
Yes — when the underlying provider offers batch pricing, we surface it as a separate column on the pricing page so you can compare online and batch side by side. The saving you see already factors in our routing discount, on top of the batch discount.
Do cached inputs cost less than fresh ones?
When the model supports it, yes. The cached-input price is shown as a separate line on each model page, and the route automatically hits cache when your request sends the same prefix as a recent call. Savings stack with our routing discount.
Are there overage fees or monthly minimums?
Solo: no minimum, no overage fee — every request is billed at the published rate, and you can stop whenever. Business plans ship monthly minimums plus committed-use discounts; ask us for a custom envelope if your team needs predictability.
How are taxes and invoices handled?
Usage is metered per request in US dollars. The console shows running totals by day, week and month, and you can export a CSV that drops straight into your finance system. Sales tax, VAT and GST are added at checkout based on the billing address you provide.