New Flagship Model Added: MiniMax M3

Date: 2026-06-07 / (1405-03-17)

Summary

We announce the addition of MiniMax's new flagship model, minimax-m3. M3 brings frontier-level coding and agentic capabilities, a 1M-token context window powered by the MiniMax Sparse Attention (MSA) architecture, and native multimodal understanding for image and video input. The model is available on v1/chat/completions, with support via v1/messages and partial support on v1/responses.

---

Details

MiniMax

MiniMax M3

minimax-m3 is MiniMax's new flagship model, combining frontier coding capability, an ultra-long 1M-token context window, and native multimodality in a single model. Built on the proprietary MiniMax Sparse Attention (MSA) architecture, M3 supports autonomous task decomposition, tool invocation, and multi-step reasoning — providing a reliable foundation for AI coding assistants and automated workflows.

Feature	Details
Context window	Up to 1,000,000 tokens (guaranteed minimum 512K)
Input pricing (≤512K)	$0.60 / 1M tokens
Cached input pricing (≤512K)	$0.12 / 1M tokens (80% cost reduction)
Output pricing (≤512K)	$2.40 / 1M tokens
Input pricing (above 512K)	$1.20 / 1M tokens
Cached input (above 512K)	$0.24 / 1M tokens
Output pricing (above 512K)	$4.80 / 1M tokens
Input modalities	Text, Image, Video
Output modalities	Text
Supported endpoints	v1/chat/completions, v1/messages (full), v1/responses (partial)

Key Features:

• Frontier Coding & Agentic Capabilities: Strong performance across software engineering and terminal execution benchmarks (SWE-Bench Pro 59.0%, Terminal-Bench 2.1 66.0%, MCP Atlas 74.2%)
• 1M Context Window: Powered by MiniMax Sparse Attention (MSA) for long-range agent tasks, long-range coding, and long-video understanding, with a guaranteed minimum of 512K tokens
• Native Multimodality: Trained with mixed-modality data from step zero, supporting image and video input for deep alignment between textual and visual semantic spaces
• Toggleable Thinking: Switch reasoning on for complex agentic and long-horizon tasks, or off for faster, latency-sensitive scenarios — both modes share the same pricing
• Autonomous Long-Horizon Tasks: Capable of multi-hour autonomous execution with tool calling and self-validation
• Context-Aware Pricing: Standard rate for the common ≤512K input range, with a higher long-context rate only above 512K tokens

Performance Highlights:

• SWE-Bench Pro: 59.0%
• Terminal-Bench 2.1: 66.0%
• SWE-fficiency: 34.8%
• KernelBench Hard: 28.8%
• MCP Atlas: 74.2%
• BrowseComp: 83.5

---

Pricing Summary

Model	Input ($/1M tokens)	Cached Input ($/1M tokens)	Output ($/1M tokens)	Above 512K (Input / Cached / Output)
minimax-m3	$0.60	$0.12	$2.40	$1.20 / $0.24 / $4.80

---

API Request/Response Examples

Example Request

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "minimax-m3",
    "messages": [
      {
        "role": "user",
        "content": "Refactor this Python function for better performance and explain your reasoning."
      }
    ],
    "max_tokens": 4096
  }'

Example Response

{
  "id": "chatcmpl-minimax-m3-abc123",
  "created": 1780270281,
  "model": "minimax-m3",
  "object": "chat.completion",
  "system_fingerprint": null,
  "choices": [
    {
      "finish_reason": "stop",
      "index": 0,
      "message": {
        "content": "Here is the refactored function with explanations...",
        "role": "assistant",
        "thinking_blocks": [],
        "annotations": []
      }
    }
  ],
  "usage": {
    "completion_tokens": 320,
    "prompt_tokens": 28,
    "total_tokens": 348,
    "completion_tokens_details": null,
    "prompt_tokens_details": {
      "audio_tokens": null,
      "cached_tokens": null,
      "text_tokens": 28,
      "image_tokens": null
    }
  },
  "estimated_cost": {
    "unit": "0.0007848000",
    "irt": 89.94,
    "exchange_rate": 114600
  }
}

---

SDK Usage Examples

Bash

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "minimax-m3",
    "messages": [
      {
        "role": "user",
        "content": "Build an agent harness that can parse a research paper and reproduce its core experiments."
      }
    ],
    "max_tokens": 8192
  }'

Python

from openai import OpenAI

client = OpenAI(api_key="your-avalai-api-key", base_url="https://api.avalai.ir/v1")

response = client.chat.completions.create(
    model="minimax-m3",
    messages=[
        {
            "role": "user",
            "content": "Build an agent harness that can parse a research paper and reproduce its core experiments.",
        }
    ],
    max_tokens=8192,
)

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

JavaScript

import { OpenAI } from "openai";

const client = new OpenAI({
  apiKey: process.env.AVALAI_API_KEY,
  baseURL: "https://api.avalai.ir/v1",
});

const response = await client.chat.completions.create({
  model: "minimax-m3",
  messages: [
    {
      role: "user",
      content: "Build an agent harness that can parse a research paper and reproduce its core experiments.",
    },
  ],
  max_tokens: 8192,
});

console.log(response.choices[0].message.content);

Anthropic SDK (v1/messages)

M3 is also available through the Anthropic Messages API format, with full support for native thinking blocks and tool use.

Bash

curl https://api.avalai.ir/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: $AVALAI_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "minimax-m3",
    "max_tokens": 4096,
    "messages": [
      {
        "role": "user",
        "content": "Design a fault-tolerant distributed task queue."
      }
    ]
  }'

Python

import anthropic

client = anthropic.Anthropic(
    api_key="your-avalai-api-key",
    base_url="https://api.avalai.ir",  # Note: without /v1
)

message = client.messages.create(
    model="minimax-m3",
    max_tokens=4096,
    messages=[
        {"role": "user", "content": "Design a fault-tolerant distributed task queue."}
    ],
)

print(message.content)

JavaScript

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  apiKey: process.env.AVALAI_API_KEY,
  baseURL: "https://api.avalai.ir",  // Note: without /v1
});

const message = await client.messages.create({
  model: "minimax-m3",
  max_tokens: 4096,
  messages: [
    { role: "user", content: "Design a fault-tolerant distributed task queue." }
  ],
});

console.log(message.content);

---

Endpoint Support

• v1/chat/completions: Full support via the OpenAI-compatible API
• v1/messages: Full support via the Anthropic Messages API format, including native thinking blocks and tool use
• v1/responses: Partial support (text input/output and basic tool use)

---

Documentation Links

• MiniMax Models Documentation
• Pricing Details
• Reasoning Models Guide
• Messages API Reference
• Responses API Reference