Pricing
You pay only for what you use.
No subscription. You buy a pack of tokens, and each run deducts the exact tokens it used — at 90% of the model provider's official price. You see the estimated cost before you run, and the tokens actually used after. Most runs cost a fraction of a cent.
What a run actually costs
Real estimates from the two most common models. Your actual cost depends on how much text you send and get back.
| Task | GPT-5 | GPT-5 |
|---|---|---|
| Polish a paragraph~250 words | $0.0033,123 tok | $0.0033,123 tok |
| Polish a full page~600 words | $0.0076,704 tok | $0.0076,704 tok |
| Read & brief a paper~5,000 words | $0.0112,569 tok | $0.0112,569 tok |
Token packs
$1 buys 1,000,000 tokens. Tokens never expire and roll over.
$5
5M tokens
$20
20M tokens
$50
50M tokens
Every new account starts with free tokens to try things out. Browse the library →
Model rates
What you pay per 1,000,000 tokens, by model (90% of the provider's official price). Output tokens usually cost more than input.
| Model | Input / 1M | Output / 1M |
|---|---|---|
| GPT-5 | $0.88 | $7.06 |
| GPT-5.2 | $1.32 | $10.58 |
| GPT-5 mini | $0.13 | $1.06 |
| Claude Sonnet | $1.69 | $8.45 |
| Claude Haiku 4.5 | $0.77 | $3.86 |
| Gemini 3 Pro | $0.88 | $5.29 |
| Gemini 3 Flash | $0.19 | $1.55 |
| DeepSeek V4 Pro | $3.31 | $6.62 |
| DeepSeek V4 Flash | $0.28 | $0.41 |