Pricing and billing
Gisoo has no subscription and no minimum plan. You hold prepaid credit in US dollars and every request is charged at the price of the model it used.
How billing works
- Prepaid credit in USD. Top up at gisoo.pro/app/credit. Top-ups are currently paid with an Iranian bank card on the NetArz payment page; NetArz is the seller and issues the receipt. The balance is shared between your Gisoo and NetArz accounts.
- A hold per request. Before a call runs, Gisoo reserves the largest amount it could cost. When the answer is finished, the hold is settled at the real cost and the rest is released.
- Real tokens, real prices. Text models are priced per million input and output tokens, and cached input is cheaper where the provider offers it. Images, audio, music and video are priced per image, per second, per minute, per request or per million characters, as shown on each model.
What is not billed
- A provider failure before any output: the hold is returned.
- Authentication, insufficient-credit, rate-limit and bad-request errors.
- A video job that fails.
GET /models,/me,/usageand/requests/{id}.
Identity verification
An account that has not completed identity verification can spend up to $2. Beyond that the API returns kyc_required until verification at gisoo.pro/app/kyc is complete.
What we keep
Prompts and responses are not stored. Billing keeps metadata only: model, token counts, cost, latency and IP address. Each request's cost record is available from GET /requests/{id} and your totals from GET /usage.
Prices of popular models
| Use | Model id | Input / 1M tokens | Output / 1M tokens |
|---|---|---|---|
| Fast and cheap chat | gpt-4o-mini |
$0.18 | $0.72 |
| Stronger general model | gpt-4.1 |
$2.80 | $11.20 |
| Reasoning | gpt-5-mini |
$0.35 | $2.80 |
| Claude, fast tier | openrouter/anthropic/claude-haiku-4.5 |
$1.5297 | $7.6487 |
| Gemini, fast tier | gemini-3.6-flash |
$1.05 | $5.25 |
| DeepSeek | deepseek-v4-flash |
$0.36 | $1.44 |
| Embeddings | text-embedding-3-small |
$0.028 | - |
Read live from the catalogue and shown in US dollars. Prices change when providers change theirs; the exact price of any model is on the models page and in GET /models.
A worked example
A request to gpt-4o-mini with 1,500 input tokens and 500 output tokens costs:
- Input: 1,500 tokens at the input price = $0.00027
- Output: 500 tokens at the output price = $0.00036
- Total: $0.00063, taken from your balance when the reply is complete.
Start
Quickstart shows how to create a key and send a request. Persian billing details: /docs/billing.