The first of the GPT-6 line: OpenAI's deepest reasoning for complex coding, computer use, research, and document creation across a 1M-token context. Reasoning effort scales all the way from low to max.
Every rate is 50% of the provider's official list price. USD per million tokens.
| Token type | List price | Tokenless |
|---|---|---|
| Input | $10.00 | $5.00 |
| Output | $50.00 | $25.00 |
| Cache write | $12.50 | $6.25 |
| Cache read | $1.00 | $0.50 |
| Prompts over 272K tokens | ||
| Input | $20.00 | $10.00 |
| Output | $75.00 | $37.50 |
| Cache write | $25.00 | $12.50 |
| Cache read | $2.00 | $1.00 |
A request whose prompt, including cached tokens, is over the threshold is billed entirely at the long-prompt rates, as the provider bills it.
Tokenless speaks the OpenAI Chat Completions format. Point your client at the base URL, set model to gpt-6-astra, and you are done.
curl https://api.tokenless.store/api/v1/chat/completions \
-H "Authorization: Bearer $TOKENLESS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-6-astra",
"messages": [{ "role": "user", "content": "Hello" }]
}'On Tokenless, GPT-6 Astra costs $5.00 per million input tokens and $25.00 per million output tokens, half of OpenAI's list price of $10.00 and $50.00. Prompts over 272K tokens bill at $10.00 input and $37.50 output per million.
Yes. Tokenless routes a request for gpt-6-astra to the genuine GPT-6 Astra and returns its response unchanged. No cheaper model is substituted.
Point any OpenAI-compatible SDK at https://api.tokenless.store/api/v1, use your Tokenless API key, and set the model to gpt-6-astra. You pay per token from a prepaid balance, and your first $1 is free.
GPT-6 Astra accepts up to 1,000,000 tokens of context and returns up to 128,000 output tokens per request.