GPT-6 Astra Python API Example

1. Install the SDK

pip install openai

2. Set your API key

export OPENAI_API_KEY="your_api_key"

3. Send a Responses API request

from openai import OpenAI

client = OpenAI()

response = client.responses.create(
    model="gpt-6-astra",
    input="Explain three practical uses of long-context reasoning."
)

print(response.output_text)

For new projects, OpenAI recommends the Responses API. Astra uses the model ID gpt-6-astra.

Use reasoning effort deliberately

response = client.responses.create(
    model="gpt-6-astra",
    reasoning={"effort": "high"},
    input="Review this migration plan and identify hidden failure modes."
)

Astra supports low, medium, high, xhigh and max. Higher effort is useful for harder multi-step tasks but can increase latency and token usage.

Production checklist

• Log request IDs and errors.
• Set explicit timeouts and retry transient failures.
• Measure token usage before sending very large prompts.
• Check whether the request crosses the 272K long-context pricing threshold.
• Pin a snapshot if reproducibility matters and a snapshot is available.