# Pricing

Source: https://docs.temprl.org/models/pricing

> How Simo is priced - pay per input token, output is free, failed and retried requests are not charged, and how to see the cost of every request.

You pay for **input tokens only**. Output is free. Every new account gets **$2 of free credit**, about 66 million input tokens.

| Model | Input | Output |
|---|---|---|
| Simo-1 Pro (`simo-1-pro`) | $0.03 per 1M tokens | Free |

The [console](https://app.temprl.org) always shows the current rates under **Billing**.

## What that means

Reading the situation is the cost; questions about it are nearly free. So ask everything you need about one input in one request.

| Request | Input tokens | Cost |
|---|---|---|
| A support ticket with 4 questions | ~250 | ~$0.0000075 |
| A one-page contract with 10 questions | ~1,500 | ~$0.000045 |
| A million short tickets | ~250M | ~$7.50 |

Images count by the tokens the model reads from them, so large screenshots cost more than a paragraph of text.

## What you are not charged for

- Requests refused before they run: a bad key, an empty balance, an invalid request.
- Failures on our side (502, 503, 504) and rate-limited requests (429). The SDK retries these automatically.

## Seeing what you spend

Every response says what it cost and what is left:

**Python**

```python
r.cost_usd       # 0.0000075
r.balance_usd    # 24.99
r.usage          # {"input_tokens": 250, "output_tokens": 0, "cost": 0.0000075}
```

**HTTP headers**

```text
x-simo-cost-usd: 0.000007500
x-simo-balance-usd: 24.999992500
```

The console's **Usage** page breaks spend down by day and by API key.

## Balance

Accounts are prepaid. Every new account gets $2 of free credit when its email is confirmed (straight away when you sign up with Google). Requests are refused with `402 insufficient_balance` once the balance reaches $0. To add more, contact us from the console's **Billing** page.