We use cookies. Cookies and browser storage keep you signed in and remember your preferences. With your permission we also measure how the site is used. You can any time. See our cookie policy.
MiniMax: MiniMax M3
minimax/minimax-m3Try MiniMax: MiniMax M3 now
Runs on your balance at the prices above. New accounts start with free usage.
One API. Every model.
Session cost $0.00
Settings
Advanced parameters
MiniMax: MiniMax M3 at a glance
- Input price is 33% lower than the median of 434 Text models.
- Context window of 1,048,576 tokens, enough for long documents and multi-turn conversations.
- Supports tool (function) calling for agents and integrations.
- Supports structured JSON output with a schema.
- Supports reasoning effort control.
- Accepts images as input (vision).
- A typical request (1,000 input / 500 output tokens) costs about $0.00315.
About MiniMax: MiniMax M3
MiniMax M3 is a multimodal foundation model from MiniMax. It is a text model in the sense that its output is text, but it accepts more than text on the way in. MiniMax positions it as a foundation model rather than a narrow single-task system, and the provider describes it as suited to long-horizon agentic work and coding, among other things.
On the input side, MiniMax M3 accepts text, images and video. On the output side it produces text only. The context window is 1,048,576 tokens, and the maximum output is 512,000 tokens. That combination of a very large context and a very large output budget is the headline capability here: you can hand the model a large body of material and still expect a substantial response rather than a short summary. Because images and video are accepted as inputs, the model can reason over visual material alongside text within the same request.
The practical use cases follow from those properties. Long-horizon agentic work is one: tasks that accumulate state over many steps and need the earlier steps to stay in view. Coding is another, particularly work that involves reading a large amount of existing source before writing anything. Video and image inputs open up tasks where the model has to describe, interpret or answer questions about visual content, or combine that content with written instructions. The large output ceiling also makes it reasonable to ask for long generated artifacts, such as extended documents or full files, in a single call.
You call MiniMax M3 through api14's OpenAI-compatible API using the model id "minimax/minimax-m3". Because the interface is OpenAI-compatible, existing OpenAI client libraries and request shapes work with a change of base URL and model id. The model supports a broad set of request parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs and top_p. Tool calling and structured outputs are both available, which matters if you are building an agent loop or need responses that conform to a schema. The reasoning and include_reasoning parameters let you control whether reasoning content is surfaced.
There is no stated limitation in the facts beyond the output ceiling of 512,000 tokens, so the main thing to check is whether your task actually needs this much context and output. If your workload is short prompts with short answers and no visual input, a smaller text model will usually be simpler to work with. MiniMax M3 makes most sense when the input is large, when it includes images or video, or when the task runs over many steps and needs tools.
How to call the MiniMax: MiniMax M3 API
- Create an API key in your dashboard.
- Install the official openai SDK for your language.
- Set baseURL to https://api14.com/api/v1 and model to "minimax/minimax-m3".
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api14.com/api/v1",
apiKey: process.env.API14_API_KEY,
});
const result = await client.chat.completions.create({
"model": "minimax/minimax-m3",
"messages": [
{
"role": "user",
"content": "Explain vector databases in two sentences."
}
]
});
console.log(result.choices[0].message.content);
Supported parameters
frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_pSpecifications
- Context length
- 1,048,576
- Max output
- 512,000
- Input
- text, image, video
- Output
- text
Requests on api14 · last 7 days
MiniMax: MiniMax M3: frequently asked questions
- How much does the MiniMax: MiniMax M3 API cost?
- MiniMax: MiniMax M3 costs Input $1.05 / 1M tokens. It is charged from your USD balance only for successful requests; failed requests are free.
- What is the context window of MiniMax: MiniMax M3?
- MiniMax: MiniMax M3 reads up to 1,048,576 tokens per request: prompt, history and documents together. It writes up to 512,000 output tokens per response.
- What can MiniMax: MiniMax M3 do?
- MiniMax: MiniMax M3 accepts text, image, video and returns text. It supports tool calling, structured JSON output, reasoning effort control.
- How do I call MiniMax: MiniMax M3 from my code?
- Point any OpenAI SDK at https://api14.com/api/v1 and set the model to "minimax/minimax-m3". No other code changes. Copy a ready example from the code panel on this page.
- What are alternatives to MiniMax: MiniMax M3?
- Comparable Text models at a similar price: Meta: Muse Glimmer 30B, Meituan: LongCat 2.0, MiniMax: MiniMax M2.7. Each runs through the same API key, so switching is one string.
MiniMax: MiniMax M3 community
New topicThis community has no topics yet. Open the first one about MiniMax: MiniMax M3.