MiniMax M3 inference overview

Price per 1M tokens

$0.23 (input)
$0.05 (cache hit)
$0.96 (output)

Parameters

23B (active)
428B (total)

Context Window

262K

Release Date

Jun 2026

MiniMax M3 inference details

MiniMax M3 is a multimodal MoE model featuring 23B activated parameters and a total of 428B parameters. It accepts text and image inputs and is optimized for long-horizon coding tasks and agentic and tool-use workflows.

Please refer to the MiniMax M3 License for terms of use. Use of MiniMax M3 or any derivatives prohibits the following:

(1) generating or disseminating content prohibited by applicable laws or regulations;
(2) assisting with or supporting any military purpose;
(3) exploiting or harming minors;
(4) generating or disseminating false or misleading information with intent to cause harm;
(5) promoting discrimination, hate speech, or harmful behavior against protected groups.

Created by: 

MiniMax

License: 

other

Model card: 

				
					import openai
import weave

# Weave autopatches OpenAI to log LLM calls to W&B
weave.init("<team>/<project>")

client = openai.OpenAI(
    # The custom base URL points to W&B Inference
    base_url='https://api.inference.wandb.ai/v1',

    # Get your API key from https://wandb.ai/authorize
    # Consider setting it in the environment as OPENAI_API_KEY instead for safety
    api_key="<your-apikey>",

    # Optional: Team and project for usage tracking
    project="<team>/<project>",
)

response = client.chat.completions.create(
    model="MiniMaxAI/MiniMax-M3",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Tell me a joke."}
    ],
)

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

MiniMax M3 resources

Screenshot 2025-07-30 at 1.03.14 PM
Course
AI engineering course: Agents
Inference_logo
Guide
W&B Inference powered by CoreWeave
Screenshot 2025-07-30 at 8.00.14 AM
Whitepaper
A primer on building successful AI agents