Pricing

Prepaid credits, metered per request. Off-peak is cheaper — most of the day, and all weekend.

Model rates

USD per 1M tokens

ModelContextInputCached inputOutput
deepseek-chat
通用对话与创作,性价比最高
1,000,000
$0.250
peak $0.425
$0.008
peak $0.014
$1.45
peak $2.46
deepseek-reasoner
深度推理,适合复杂任务与代码重构
1,000,000
$1.10
peak $1.87
$0.040
peak $0.068
$4.75
peak $8.07
qwen-plus
通义千问,国内调用延迟低
131,072
$0.400
$0.100
$1.20

Upper line = off-peak (all weekend, and weekdays outside 01:00-04:00 / 06:00-10:00 UTC). Lower line = peak rate.

How billing works

Peak and off-peak rates

Model providers charge more during their peak windows (UTC 01:00-04:00 and 06:00-10:00, Monday to Friday). We pass through a portion of that: off-peak is roughly 40% cheaper. Your balance is always charged at the rate in effect when the request arrives.

Metered, not estimated

Charges come from the token counts reported by the provider on each response. Where a provider omits usage, the request is marked estimated and audited.

Cached input is cheaper

When the provider reports a prompt cache hit, those tokens bill at the cached rate instead of the input rate.

Failed requests are free

If a request fails before producing output, the pre-authorised amount is released and nothing is deducted.

No expiry, no auto-recharge

Credits stay on your balance indefinitely. Your card is never charged without an explicit purchase.

Subscriptions

Monthly credits at a discount. Optional — pay-as-you-go works without a plan.

Starter
$20.00/mo
$2.00 in monthly credits
Pro
$100.00/mo
$11.00 in monthly credits · 5% off usage
Business
$500.00/mo
$57.50 in monthly credits · 10% off usage

Top-up amounts

starter
$5.00
$5.00 in credits
pro
$20.00
$20.00 in credits · +$0.100 bonus
business
$100.00
$100.00 in credits · +$1.00 bonus
Create an accountRead the API docs

Prices are in US dollars and exclude taxes that may apply in your jurisdiction. EU business customers may provide a VAT number at checkout for reverse charge.