Free trial: new accounts receive trial credits, no card required — enough to run a few thousand requests and measure real latency from your region before spending anything.
Per 1 million tokens, in USD. Our price is the provider's cost plus a flat markup, so what you pay tracks what we pay.
Illustrative figures. Upstream prices change often and vary by provider tier. Replace every number below with live values from your upstream dashboards, then delete this box and add a "last updated" date.
| Model | Input / 1M | Output / 1M | Billing |
|---|---|---|---|
deepseek-chat | $0.33 | $1.33 | per 1M tokens |
deepseek-reasoner | $0.67 | $2.67 | per 1M tokens |
qwen-plus | $0.13 | $0.33 | per 1M tokens |
qwen-max | $0.83 | $3.33 | per 1M tokens |
glm-4.6 | $0.40 | $1.60 | per 1M tokens |
kimi-k2 | $0.60 | $2.40 | per 1M tokens |
doubao-pro | $0.22 | $0.67 | per 1M tokens |
hunyuan-turbo | $0.20 | $0.60 | per 1M tokens |
Cached input tokens, where the provider supports them, are billed at the provider's discounted rate plus the same markup.
Charged for exact input and output tokens reported by the provider. No rounding to blocks, no per-request surcharge.
Top up any amount. Usage is deducted as you go and your balance never expires.
Set a hard monthly limit on each key. It stops rather than surprising you.
Cost per model, per key, per day. Exportable for your own finance tooling.
Visa, Mastercard and other major cards. We never see or store card numbers.
One-tap top up where your browser supports it.
For teams without international cards, or who prefer to keep spend off a bank statement.
All amounts are charged in USD. Your bank may apply its own conversion rate and fees.