Pricing
You pay per million tokens, with no subscription. Input tokens are what you send, output tokens are what Kalle writes.
Prices per 1M tokens
| Model | Model ID | Input | Output |
|---|---|---|---|
| Kalle Turbo | kalle-turbo | $0.05 | $0.25 |
| Kalle Design | kalle-design | $1 | $5 |
| Kalle Turbo Max | kalle-turbo-max | $1 | $5 |
| Kalle Turbo Ultra | kalle-turbo-ultra | $2 | $10 |
With auto, you pay the price of the model that answered. The response's model field tells you which one that was.
What counts as a token
- Input: the system prompt, every message in the conversation, tool definitions, tool results, images and documents.
- Output: the text Kalle writes, its tool calls, and the thinking it does at higher effort, shown as
usage.reasoning_tokens.
Every response has a usage object with the exact numbers, so you can track the cost of each request.
Worked examples
A request with 2,000 input tokens and 500 output tokens costs:
| Model | Input | Output | Total |
|---|---|---|---|
| Kalle Turbo | $0.0001 | $0.000125 | $0.000225 |
| Kalle Design | $0.002 | $0.0025 | $0.0045 |
| Kalle Turbo Max | $0.002 | $0.0025 | $0.0045 |
| Kalle Turbo Ultra | $0.004 | $0.005 | $0.009 |
Keeping costs down
- Use Kalle Turbo for high-volume work and move up only where the answers need it.
- Keep
effortatautoorlowfor simple tasks. Thinking is billed as output. - Summarise long conversations instead of sending the full history every time.
- Crop images to what matters and send only the pages of a document you need.
Usage and spend per key and per model are shown in Kalle API in the Appmake console. For Appmake plans and credits, see Pricing.