GPT-6 Astra · Cached input · 20M tokens

Twenty million cached input tokens on GPT-6 Astra at $1 per million - 10% of its $10 input rate. A cache hit is what keeps a repeated prompt cheap.

What you get

• 20M tokens of cached input tokens on one API key. Order several units to stack tokens on the same key.
• A personal API key for gpt-6-astra, issued automatically once your order is completed.
• Priced at $1.00 per million cached input tokens. Source: OpenAI's own pricing page.
• 1,050,000 context · 128,000 max output · reasoning effort: low / medium / high / xhigh / max · vision: Yes
• No shipping: delivery is digital, under My Account > Order History and by email.

Capabilities

• 10% of input price
• Cache reads
• 1.05M context
• Vision

Best for

A long system prompt or document read on every call.

Quick start

# pip install openai   (the sample below calls the endpoint directly)

import os, requests

resp = requests.post(
    "https://api.openai.com/v1/responses",
    headers={
        "Authorization": "Bearer " + os.environ["OPENAI_API_KEY"],
        "Content-Type": "application/json",
    },
    json={
        "model": "gpt-6-astra",
        "input": "Write a one-sentence bedtime story about a unicorn.",
    },
)
print(resp.json()["output"][0]["content"][0]["text"])