Skip to content
Router One
Sign up

LLM API Pricing and Subscription Plans

Router One pricing has two parts: wallet pay-as-you-go billed per token at the rates posted on /models (up to 90% off official prices on select models), and Pro, Max and Ultra monthly subscriptions (from $20.19/month) with a metered model allowance — usage outside plan coverage draws on the wallet.

Two ways to pay: the monthly subscription plans below, or wallet pay-as-you-go billed per token. Jump to pay-as-you-go

Pro

Developer favorite

$20.19/ month

Model allowance

30-day cycle

Standard models

5,000 calls / cycle
View 7 included models
claude-haiku-4.5gemini-3-flash-previewgemini-3.8-flashgemini-3.7-flashgemini-3.6-flashgemini-3.5-flashgemini-3.5-flash-lite

Mid-tier models

1,200 calls / cycle
View 5 included models
claude-sonnet-5claude-sonnet-4.6gpt-5.6-terragrok-4.5grok-4.6

Premium models

300 calls / cycle
View 8 included models
gpt-5.6-solclaude-opus-4.8claude-opus-4.7-thinkingclaude-opus-4.6claude-opus-5gpt-6.1-solgpt-6-solgemini-4-argon

Flagship models

—

Usage beyond these allowances is billed to your wallet balance.

Card auto-renewal · Wallet balance can pay one 30-day period

Max

Power user standard

Recommended
$47.99/ month

Model allowance

30-day cycle

Standard models

8,000 calls / cycle
View 7 included models
claude-haiku-4.5gemini-3-flash-previewgemini-3.8-flashgemini-3.7-flashgemini-3.6-flashgemini-3.5-flashgemini-3.5-flash-lite

Mid-tier models

4,500 calls / cycle
View 5 included models
claude-sonnet-5claude-sonnet-4.6gpt-5.6-terragrok-4.5grok-4.6

Premium models

1,000 calls / cycle
View 8 included models
gpt-5.6-solclaude-opus-4.8claude-opus-4.7-thinkingclaude-opus-4.6claude-opus-5gpt-6.1-solgpt-6-solgemini-4-argon

Flagship models

150 calls / cycle
View 1 included model
gpt-6-astra

Usage beyond these allowances is billed to your wallet balance.

Card auto-renewal · Wallet balance can pay one 30-day period

Ultra

Enterprise / SaaS platform

$149.00/ month

Model allowance

30-day cycle

Standard models

25,000 calls / cycle
View 7 included models
claude-haiku-4.5gemini-3-flash-previewgemini-3.5-flashgemini-3.6-flashgemini-3.7-flashgemini-3.8-flashgemini-3.5-flash-lite

Mid-tier models

15,000 calls / cycle
View 5 included models
claude-sonnet-5claude-sonnet-4.6gpt-5.6-terragrok-4.5grok-4.6

Premium models

5,000 calls / cycle
View 8 included models
gpt-5.6-solclaude-opus-4.8claude-opus-4.7-thinkingclaude-opus-4.6claude-opus-5gpt-6.1-solgpt-6-solgemini-4-argon

Flagship models

450 calls / cycle
View 1 included model
gpt-6-astra

Usage beyond these allowances is billed to your wallet balance.

Card auto-renewal · Wallet balance can pay one 30-day period

Compare plan allowances

Allowance per 30-day cycle (a standard request uses 1; long-context requests use more). Models in a tier share that tier's allowance.

Model tierPro$20.19 / monthMax$47.99 / monthUltra$149.00 / month
Standard models5,0008,00025,000
Mid-tier models1,2004,50015,000
Premium models3001,0005,000
Flagship modelsNot included150450

Prefer no subscription? Top up your wallet and pay only for what you use. See how pay-as-you-go billing works

Top up to get started

How long-context calls use your quota

A standard request uses 1 request from its model tier. A request whose input exceeds a model’s long-context threshold uses more, by input-length band, as listed below. Model coverage depends on your plan.

Input ≤ 256,000 tokens
1 request
256,000 < input ≤ 512,000 tokens
2 requests
Input > 512,000 tokens
4 requests
Input ≤ 272,000 tokens
1 request
272,000 < input ≤ 512,000 tokens
2 requests
Input > 512,000 tokens
4 requests
Input ≤ 272,000 tokens
1 request
Input > 272,000 tokens
2 requests
Input ≤ 128,000 tokens
1 request
128,000 < input ≤ 256,000 tokens
2 requests
256,000 < input ≤ 512,000 tokens
4 requests
Input > 512,000 tokens
8 requests
Input ≤ 200,000 tokens
1 request
Input > 200,000 tokens
2 requests

Thresholds use the total input of one request, including cached input without double-counting. Output does not affect the threshold. Input exactly at a threshold stays in the lower tier. If quota is insufficient, your plan’s existing overage rules apply; calls billed to the wallet use the posted token rates. Browse model rates

Pay-as-you-go LLM API pricing

Prefer usage-based billing? Top up your wallet and pay each model's posted input, output, cached-input, or per-generation rate. This is separate from the monthly subscription allowances above.

Across 59 priced models, standard-context input and output rates run from $0.06 to $40.00 per 1M tokens. Cached-input rates and long-context rate lines, where a model has them, are listed separately on its model page, and the model catalog is the source of truth for every rate.

Top up to get started
Wallet pay-as-you-go
Top up prepaid USD wallet credits with a card or Alipay through one hosted checkout, or with USDT/USDC stablecoins. API calls deduct from balance by token usage.
Posted model rates
The /models page is the source of truth for input, output, cached input, and unit-priced model rates. Dashboard → Logs shows what each call cost, with any account or model discount and plan multiplier already applied.
How far below list price
Pricing starts as low as 10% of official provider list prices (up to 90% off) on select models. Discount levels vary by model — the model catalog marks each model's discount, and a model's detail page shows the discounted rate against its base rate.
Subscriptions alongside the wallet
Pro, Max, and Ultra subscriptions are available for predictable monthly access. Wallet balance remains available for one-time periods and usage outside plan coverage. Read the subscription launch post
Visible checkout fees
Currency conversion, payment-channel fees, and stablecoin network assumptions are shown before checkout. There is no hidden markup inside the pay-as-you-go token line.

Billing rules

Subscription quota is separate from wallet balance
Subscriptions provide request-based monthly quotas for eligible model tiers. Wallet balance is a separate pay-as-you-go pool billed by actual token usage. Using Claude Code? Compare with Claude Max
Shared monthly request quotas
Models in the same tier share that tier's monthly request allowance; the quota is not granted separately per model. Tier assignments follow what each plan card displays.
After the request quota is used
Additional eligible calls and any usage outside plan coverage draw from your wallet balance at the posted pay-as-you-go rate.
Per-token billing on pay-as-you-go
Pay-as-you-go calls are billed per token — input, output, and cached input each at the model's posted rate. See exact rates on the /models page. Buying for many customers? How volume spend works

Payment methods

MethodDetailsInvoice
CardInternational credit and debit cards through one hosted checkout, billed in USD. Recommended for teams outside mainland China.Automatic
AlipayThe same hosted checkout also offers Alipay, so a top-up settles in RMB without a US credit card. Works for individuals and small teams in mainland China.Automatic
USDT / USDCTop up with USDT or USDC on Tron, BSC, Ethereum, Polygon, Base, or Arbitrum. Useful for global teams that prefer stablecoin treasury workflows or non-card billing.Not invoiced

Refunds, balance, and invoicing

Monthly request quota
Unused subscription requests do not roll over. The plan quota resets at the end of each billing cycle, while wallet balance stays separate.
Wallet balance
Wallet top-ups never expire. Balance can be spent on any model at posted pay-as-you-go rates.
Refund policy
Purchased credits are non-refundable once added to your balance, except for billing errors on our side or where required by law; bonus credits are non-refundable too. We recommend a small first top-up to try the service. Read the refund policy
Invoices
Every payment completed through the hosted checkout — card or Alipay — is invoiced automatically: the invoice is emailed to the address you entered at checkout as soon as the payment succeeds, with nothing to request. Plain pay-as-you-go wallet top-ups get one too, not only enterprise or subscription accounts. Stablecoin top-ups are a separate rail and are not invoiced.

Frequently asked questions

Will unused subscription requests carry over to the next billing cycle?
No. Each plan's request quota resets at the end of its billing cycle. If your usage is bursty, keep wallet balance as a buffer.
Can I use a subscription and pay-as-you-go at the same time?
Yes. Subscription coverage and wallet pay-per-token billing can work together, with wallet balance covering usage outside plan coverage.
Can I get an invoice?
Yes. Every payment completed through the hosted checkout (card or Alipay) is invoiced automatically — the invoice is emailed to the address you entered at checkout once the payment succeeds, for pay-as-you-go wallet top-ups and subscriptions alike, so there is nothing to request. Stablecoin top-ups are a separate rail and are not invoiced. For annual contracts, contact sales through the /contact page.
What is the minimum top-up amount?
The checkout page is the source of truth for the minimum amount. Card and Alipay top-ups currently usually start at $5 USD equivalent; RMB amounts are converted and shown at checkout using the current FX rate. Stablecoin top-ups (USDT/USDC on Tron, BSC, Ethereum, Polygon, Base, Arbitrum) usually start at $5 USD equivalent; network gas is paid by the user and any channel fee is shown before checkout.
Does Router One retain my request or response data?
Router One does not retain prompt or model-response bodies from direct API calls. It records request metadata for billing, usage tracking, and troubleshooting. Playground saves conversation history so users can revisit and continue their chats.
How is billing handled when a provider retry succeeds?
A named-model request is billed at that model's posted rate when an eligible retry succeeds on another provider route. With model="auto", billing follows the final server-selected model. Customer Logs shows the model, request ID, and settled request cost; upstream suppliers, service routes, and intermediate attempts remain internal.
Can users in mainland China call international models?
Yes — Router One proxies the request through a China-reachable endpoint, so models hosted overseas work without a VPN.
Where can I see exact per-model prices?
The /models page lists every supported model with its applicable token or per-unit price, plus context length and capabilities where relevant.
How is a subscription different from adding the same amount to my wallet?
A subscription covers eligible models through shared monthly request quotas: standard requests use 1 request, while requests above a model’s long-context threshold use more, by input-length band. Wallet balance is billed per token and never expires. You can use both, with the plan’s overage rules applying when quota is insufficient. Bonus credits from deposits are spendable on API usage only and do not pay for a plan.
Which payment methods can I actually use today?
Two rails. One hosted checkout takes major credit and debit cards and Alipay — pick either after entering your billing details. The other rail is USDT/USDC on Tron, BSC, Ethereum, Polygon, Base, or Arbitrum. WeChat Pay is not currently supported; use Alipay for an RMB top-up without a US credit card.
Do long-context requests use more of my subscription quota?
Yes. A standard request uses 1 request from its model tier; a request above a model’s long-context threshold uses more, by input-length band. Available quota rules appear below the plans; model detail pages also list input-length tiers. Only total input strictly above a threshold enters the next tier. Cached input is included once; output does not affect the threshold. If quota is insufficient, the plan’s overage rules apply. Wallet-billed calls use the model’s posted token rates.

Ready to route your first request?

One API key for 30+ models, with per-request cost visibility. Top up, go live in minutes, and pay only for what you use.

Top up to get startedBrowse model rates