Token pay-as-you-go deployment

​

Overview

Billed by Token usage (no charge when not in use). Supports only some models that have undergone LoRA efficient fine-tuning. Suitable for post-tuning model effect validation and low-cost scenarios with low concurrency and latency requirements. Throughput/concurrency and generation speed are preset by the platform and cannot be adjusted by users.

​

Billing rules

Fee = Model input token count × Model input unit price + Model output token count × Model output unit price (minimum billing unit: 1 token)

• Billing by model token usage is supported only after you complete efficient SFT training (that is, LoRA efficient fine-tuning; the plan parameter is set to lora for API deployment) on the following base models and obtain a custom model.

​

Supported models and pricing

​

LoRA deployment

Token pay-as-you-go deployment supports only LoRA fine-tuned models. When creating via API, set plan to lora. The capacity parameter has no effect but must be filled in. For scaling, go to the Dedicated Deployment console and submit a form application. For a complete example of creating a deployment via API, see the API deployment guide.

​

Scaling

For Token usage-based deployments, scaling requires submitting an application form in the console and waiting for manual review. Self-service scaling is not supported.

​

FAQ

Q: What happens if not used for a month?

Token usage-based deployments will be automatically released if not used within one month.

Q: How do I switch to another billing method?

You must take the original resource offline and then create a new resource using the desired billing method. See Introduction to model deployment.