Models & Pricing
How billing works
Pay-as-you-go uses official USD rates, billed per token in real time:
- Each model has separate input and output prices, typically per 1M tokens.
- Cost of a request = input tokens × input price + output tokens × output price (some models also bill cache read/write, images, etc.).
- Different groups/plans may settle at different rates — your plan determines what you pay.
💡 In short: you pay for what you use, same logic as the official APIs — just settled through HJToken.
How image models are billed
Image generation (gpt-image-*) is billed per output token, not per image: higher quality and larger size produce more output tokens, so they cost more.
On the pricing page, "Output (/1M)" is the text output rate — for images, see the "Image output (/1M)" column.
See Image Generation for details.
Where to see live prices
Models and prices change upstream, so they are shown on the site (auto-synced, no need to read docs):
- 📊 Pricing · hjtoken.cn/model-pricing — per-model base prices
- 🔎 Available Channels · hjtoken.cn/available-channels — platforms/models your account can access
Currency & exchange rate
- Prices are USD-based.
- When paying via a CNY channel (e.g. Alipay), the amount is converted at the current exchange rate. The checkout page shows the rate and the final amount — what you see is what you pay.
Pay-as-you-go vs subscription
| Your case | Suggestion |
|---|---|
| Light / bursty usage | Pay-as-you-go (start with the trial credit you claim from support) |
| Steady heavy usage | Subscription (VIP / SVIP) — lower unit cost, see Subscriptions |
Track usage: Usage · hjtoken.cn/usage.