LLM Reference
MiniMax

Using MiniMax M3 on MiniMax

Implementation guide · MiniMax M3 · MiniMax

Serverless

Quick Start

  1. 1
    Create an account at MiniMax and generate an API key.
  2. 2
    Use the MiniMax SDK or REST API to call MiniMax-M3 — see the documentation for request format.
  3. 3
    You'll be billed $0.60/1M input, $2.40/1M output tokens.
API PortalDocumentationModel Card

Code Examples

See MiniMax documentation for integration details.

About MiniMax

MiniMax is a multimodal foundation model and API platform for text, speech, video, image, and music generation with agent tools.

View all models on MiniMax →

Pricing on MiniMax

TypePrice (per 1M)
Input tokens$0.60
Output tokens$2.40
Image input$1.00
Video input$1.00

Capabilities

VisionMultimodalReasoningFunction CallingTool UsePrompt Caching

About MiniMax M3

MiniMax M3 is a frontier multimodal model released May 31, 2026, accepting text, image, and video inputs with up to 1 million tokens of context. Built on MiniMax Sparse Attention, it targets production coding and long-horizon agent workflows with lower long-context attention cost, 9.7x faster prefill, 15.6x faster decode versus MiniMax M2 at 1M tokens, interleaved reasoning between tool calls, function calling, and automatic prompt caching.

Full model details →

Model Specs

Released2026-05-31
Context1m
ArchitectureMiniMax Sparse Attention (MSA)

Provider

MiniMax
MiniMax