Billing and usage

How billing works#

• Text models: input and output tokens are priced separately; tokens served from the prompt cache are billed at the lower cache price.
• Reasoning: thinking tokens are billed at the output price.
• Image models: billed per image, with the price depending on resolution.
• Each model's price is on the model pricing page, and the listed price is exactly what you're charged.

An estimated amount is held when a request starts and settled against actual usage when it finishes. Requests that fail without producing output are refunded in full.

Adding funds#

Top up online on the console's top-up page with Alipay or WeChat Pay; funds are available immediately. You can also redeem a top-up code.

Viewing usage#

The console's usage log lists every call with its time, model, input and output tokens, cache hits and cost. You can filter by key, model and time.

Checking the balance with the API#

You can query a key's budget and usage with the key itself, which is handy for scripts and monitoring. See Balance and usage.

curl https://api.flyrouter.ai/v1/dashboard/billing/subscription \
  -H "Authorization: Bearer $FLYROUTER_API_KEY"

Keeping costs down#

• Create a separate key for each project or tool and give it a spending limit. See Authentication.
• Turn on prompt caching for long system prompts and reference documents.
• Use lightweight models for simple tasks, or lower the reasoning effort.
• Resize images to the resolution you need before sending them.