Gemini Image
The Chinese site currently advertises reference image pricing for selected models. Confirm the live model card and console before budgeting a batch.
English home / Pricing
Pricing and billing
Pay for the work you actually run
TokenAI uses usage-based billing. Image requests are commonly billed per image, video jobs by the model's task or duration rules, and text models by token usage. Always verify the live dashboard before production.
How to read the bill
| Workload | Typical unit | What to record |
|---|---|---|
| Text and reasoning | Input/output tokens | Model ID, usage fields, request ID and balance delta |
| Image generation | Per image or model-specific unit | Image count, size/quality options and final ledger entry |
| Video generation | Task, duration or model-specific rule | Task ID, terminal status, duration and final charge |
A failed request and a repeated request are different events. Save the request or task ID before retrying so you can reconcile the ledger.
Reference price examples
Seedance 2.5
A limited campaign may advertise a reference price for one second of output. Seedance runs on a separate site, and its account and billing do not carry over from the main site.
Text models
Token pricing can vary by model, context tier and routing group. Do not infer a fixed USD conversion from an old screenshot.
Payment and support
Do not send funds to an address found in a message. Check the live dashboard for supported payment methods. USDT settlement is not advertised here until the actual payment flow is live.
For account or billing questions, join the Telegram group with a redacted request ID.