Skip to content

Multi-model API gateway

Top up once. Pay as you go.

Prepaid USD balance, metered at published token rates as you go.

Billing
Prepaid USD balance
Billing
One-time purchase · no auto-renew
Expiry
Balance never expires

One-time recharge tiers

$10.00USD

You pay$10

$20.00USD

You pay$20

$50.00USD

You pay$50

$100.00USD

You pay$100

$500.00USD

You pay$500

Purchases are processed by Waffo Pancake as merchant of record. By buying, you agree to our Terms and acknowledge our Refund Policy, Privacy Policy, and Acceptable Use Policy.

  1. 1Pick a tier and pay
  2. 2Balance lands instantly
  3. 3Metered as you go

Published model pricing

deepseek-v4-1-flash-1017

1000K ctx · Streaming

InputUSD / 1M
Peak$0.150
Off-peak$0.088
OutputUSD / 1M
Peak$0.599
Off-peak$0.351
Cache hitUSD / 1M
Peak$0.0039
Off-peak$0.0039

deepseek-v4-flash

1000K ctx · Streaming · 9% below list

InputUSD / 1M
$0.136
OutputUSD / 1M
$0.273
Cache hitUSD / 1M
$0.027

deepseek-v4-flash-0731

1000K ctx · Streaming · 22% below list

InputUSD / 1M
Peak$0.175
Off-peak$0.146
OutputUSD / 1M
Peak$0.526
Off-peak$0.438
Cache hitUSD / 1M
Peak$0.0058
Off-peak$0.0049

deepseek-v4-pro

1000K ctx · Streaming · 35% below list

InputUSD / 1M
$1.17
OutputUSD / 1M
$2.34
Cache hitUSD / 1M
$0.097

deepseek-v4-pro-0813

1000K ctx · Streaming · 22% below list

InputUSD / 1M
$0.526
OutputUSD / 1M
$1.58
Cache hitUSD / 1M
$0.018

deepseek-v3-2

164K ctx · Streaming

InputUSD / 1M
$0.779
OutputUSD / 1M
$1.17
Cache hitUSD / 1M
$0.078

glm-5-3-flash

1000K ctx · Streaming · 16% below list

InputUSD / 1M
$0.101
OutputUSD / 1M
$0.354
Cache hitUSD / 1M
$0.029

glm-5-3

1000K ctx · Streaming · 35% below list

InputUSD / 1M
$0.779
OutputUSD / 1M
$2.73
Cache hitUSD / 1M
$0.195

glm-5-2

1000K ctx · Streaming · 35% below list

InputUSD / 1M
$0.779
OutputUSD / 1M
$2.73
Cache hitUSD / 1M
$0.195

glm-5-1

200K ctx · Streaming · 28% below list

InputUSD / 1M
$0.643
OutputUSD / 1M
$2.57
Cache hitUSD / 1M
$0.139

glm-5

200K ctx · Streaming · 28% below list

InputUSD / 1M
$0.429
OutputUSD / 1M
$1.93
Cache hitUSD / 1M
$0.107

kimi-k3

1000K ctx · Streaming · 16% below list

InputUSD / 1M
$2.53
OutputUSD / 1M
$12.66
Cache hitUSD / 1M
$0.253

kimi-k2-7-code

256K ctx · Streaming · 35% below list

InputUSD / 1M
$0.633
OutputUSD / 1M
$2.63
Cache hitUSD / 1M
$0.253

kimi-k2-6

256K ctx · Streaming · 22% below list

InputUSD / 1M
$0.760
OutputUSD / 1M
$3.16
Cache hitUSD / 1M
$0.214

qwen-3-7-max

1000K ctx · Streaming · 51% below list

InputUSD / 1M
$0.877
OutputUSD / 1M
$2.63
Cache hitUSD / 1M
$0.467

ernie-5-1

128K ctx · Streaming · 54% below list

InputUSD / 1M
$0.273
OutputUSD / 1M
$1.23
Cache hitUSD / 1M
—

longcat-2-0

1000K ctx · Streaming · 48% below list

InputUSD / 1M
$0.390
OutputUSD / 1M
$1.56
Cache hitUSD / 1M
$0.0078

Frequently asked questions

Common questions about balance, top-ups, and billing. Can't find an answer? Email us.

Contact support
How is usage billed?

Every API call is metered in real time at the published token rate for the model route you use. The cost is deducted from your prepaid USD balance and recorded in a usage ledger. Rates already include the Relay platform service fee; applicable taxes are shown at checkout.

How does balance work?

Recharge tiers come in one-time $10–$500 and ¥10–¥500 options. The credited USD balance never expires while your account remains available.

What happens when balance runs out?

When your balance is insufficient for the estimated cost of a request, the gateway pauses that call. Top up again, or contact support for higher volume.

Do I need a subscription?

No. Relay is pure prepaid: buy a recharge tier when you need balance, and nothing renews automatically.

Who processes payments?

Waffo Pancake acts as merchant of record for recharge purchases. Its checkout handles payment, applicable indirect taxes, and invoices. Your receipt or card statement may reference Waffo.

How do I view usage and costs?

Sign in to the dashboard to see request volume, token consumption, latency, and cost breakdowns for reconciliation and budgeting.

How is API request data handled?

Relay forwards request content to the backend model route selected for your call so a response can be generated. We record billing and diagnostic metadata. Provider dependencies and data handling are explained in the Privacy Policy.

Can I request a refund?

Eligible first purchases, duplicate charges, and delivery problems are covered by our published Refund Policy. Email [email protected] with your order or invoice number. Consumed balance is generally non-refundable.