Together AIIntegrationPasteChecked against the official docs, Nov 9, 2026

Chat with an open model (MiniMax M3.1 Flash) on Together AI

Send a prompt to the open MiniMax M3.1 Flash model on Together AI's serverless API from Python and print the reply.

Your keys

0 of 1 ready

Create each key on its own page, paste it below, then copy the finished lines into your project's .env file. What you paste stays in this tab. It is never sent to Wirestack or saved.

Together AI API keyTOGETHER_API_KEYGet key
Select Create key on your project's API keys page. The key is shown only once.

Save them once on your computer, and every project can use them:npx wirestack keys add together-aithennpx wirestack keys fill

Set up with one command

$npx wirestack add chat-open-model-together-ai-python@2

Revision 2, checked Nov 9, 2026. Installs exactly together 3.1.1, creates the files and adds the key names to your .env. It shows everything first and asks before changing anything.

Adapt to my stackProShowing the original (Python). Pick your stack to see it rewritten.

Free accounts can open one adapted integration a month. Sign in to try it.

3 of 3 free copies left this month.Unlimited with Pro

What you get

A short Python script that sends one question to MiniMax M3.1 Flash, an open model that Together AI runs on its serverless API, and prints the answer. You pay per token and there is no server to set up.

Prerequisites

Python 3.10 or newer (together 3.1.1 declares requires_python >=3.10).
A Together AI API key in TOGETHER_API_KEY. Create one on your project's API keys page: https://api.together.ai/settings/projects/~current/api-keys (the key is shown only once, so copy it).
Credits on your Together account. There is no free trial: the first purchase is at least $15, and API access stops when the balance hits zero. Add credits at https://api.together.ai/settings/organization/~current/billing

Steps

Install the SDK:

Shell

pip install together==3.1.1

Set your key:

Shell

export TOGETHER_API_KEY="your_api_key"

On Windows PowerShell: $env:TOGETHER_API_KEY="your_api_key"

Save this as main.py (copied from the Together quickstart):

Python

from together import Together

client = Together()  # reads TOGETHER_API_KEY from environment

response = client.chat.completions.create(
    model="MiniMaxAI/MiniMax-M3.1-Flash",
    messages=[
        {
            "role": "user",
            "content": "What are the top 3 things to do in New York?",
        }
    ],
)

print(response.choices[0].message.content)

The client reads TOGETHER_API_KEY on its own. Pass api_key= to Together() if you want to set it in code instead.

Test it

Shell

python main.py

After a few seconds the answer prints in your terminal. A 429 or 503 error means Together is busy: wait and retry with backoff. If calls fail with a working key, check that your credit balance is above zero.

Not confirmed

MiniMax M3.1 Flash is a reasoning model. The quickstart says its thinking comes back in a separate field when streaming; it does not say whether message.content can include thinking text in this non-streaming call.
The SDK README says Python 3.9+, but the PyPI metadata for 3.1.1 says 3.10+. We went with PyPI.
SDK 3.0.0 came out on 22 October 2026 with breaking changes (no /v1 in the base URL, beta paths removed). 3.1.0 (3 November) and 3.1.1 (4 November) only changed clusters, fine-tuning, secrets and RL tools, not chat completions. The quickstart code above is unaffected as far as the docs show, but we have not run it.
Together publishes no fixed rate limits for serverless; it may throttle at high demand.
The version pin is ours; the quickstart's install line is just pip install together.