Skip to content
TokenGauge
Menu

Setup note · 3 steps

SayGM in OpenAI SDK

SayGM accepts OpenAI-format requests. Existing code, prompts and response parsing stay the same.

Updated 9 Oct 2026

§ 01

Steps

  1. Step 1: Install the openai package for your language.

  2. Step 2: Construct the client with base_url set to the SayGM endpoint and api_key set to your SayGM key.

  3. Step 3: Pass the model ID exactly as listed on the SayGM models page.

Python
import osfrom openai import OpenAI client = OpenAI(    base_url=os.environ["SAYGM_BASE_URL"],    api_key=os.environ["SAYGM_API_KEY"],) response = client.chat.completions.create(    model="gpt-5.5",    messages=[{"role": "user", "content": "Explain TEE attestation in one paragraph."}],)

§ 02

Why route OpenAI SDK through a gateway

  • 01

    No code changes beyond the client constructor.

  • 02

    GPT-5.5 and GPT-5.6 are about 29 percent below OpenRouter in the October snapshot.

  • 03

    Gateway privacy backed by published hardware attestation.

§ 03

Models worth trying in OpenAI SDK

Snapshot SayGM prices per million tokens, input then output.

§ 04

OpenAI SDK questions

Does OpenAI SDK work with SayGM?

Yes. SayGM accepts OpenAI-format requests. Existing code, prompts and response parsing stay the same.

Which models can I use in OpenAI SDK through SayGM?

Any model in the SayGM catalogue, including Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, Claude Opus 5 and confidential open-weight models.

Where do I find the SayGM base URL?

The endpoints for the OpenAI, Anthropic and Gemini compatible APIs are listed in the SayGM developer docs, linked from the SayGM website after sign-up.