MiniMax

FeaturedAPI-first

Frontier AI platform with M3 — 1M context, coding & agents

AI Coding Assistants platform.minimax.io GitHub · 3k ★ Web API

Coding Text / Chat Image Input Image Gen Video Gen Audio Gen Agents Web Search

Try MiniMax

About

MiniMax offers frontier AI models (M3, M2.7) optimized for coding, agentic workflows, and multimodal tasks, plus MiniMax Code — a dedicated agent coding harness available on desktop and web. MiniMax M3 features a 1M token context window, strong coding benchmarks, and native multimodal support across text, image, video, and audio in a single architecture.

Key Features

1M token context window (M3)MiniMax Code agent harness (desktop & web)Agentic workflows & tool useNative multimodal (text, image, video, audio)Function callingStreaming

Pros

1M token context window on the flagship M3 model
Strong coding and agentic benchmark performance
Native multimodal — text, image, video, and audio in one architecture
Competitive API pricing from $0.30/1M tokens
MiniMax Code agent harness bundles chat, memory, and scheduling in one place

Cons

Chinese company may raise data residency/privacy concerns for some enterprises
Smaller developer community and ecosystem than OpenAI or Anthropic
MiniMax Code (the agent/IDE product) is newer and less proven than established coding tools
Fewer third-party IDE integrations than Cursor or GitHub Copilot

Pricing Plans

API

Pay-per-token from $0.30/1M tokens

Free

Get Started

MiniMax M3 & M2.7 access
Pay per token usage
1M token context window
Function calling & tool use

Go

M Plan – for individual developers

$22/mo

Start Go

Everything in API
3-4 concurrent agents
Priority access
Higher rate limits
Email support

Explore

M Plan – for power users

$54/mo

Start Explore

Everything in Go
4-5 concurrent agents
Highest rate limits
Early model access
Priority support

Build

M Plan – for teams & heavy use

$130/mo

Start Build

Everything in Explore
6-7 concurrent agents
Team management
Custom rate limits
Dedicated support

API

Pay-per-token from $0.30/1M tokens

Free

Get Started

MiniMax M3 & M2.7 access
Pay per token usage
1M token context window
Function calling & tool use

Go

M Plan – for individual developers

$22/mo

Start Go

Everything in API
3-4 concurrent agents
Priority access
Higher rate limits
Email support

Explore

M Plan – for power users

$54/mo

Start Explore

Everything in Go
4-5 concurrent agents
Highest rate limits
Early model access
Priority support

Build

M Plan – for teams & heavy use

$130/mo

Start Build

Everything in Explore
6-7 concurrent agents
Team management
Custom rate limits
Dedicated support

1–3 of 4

Supported AI Models

MiniMax M3

MiniMax / 1M ctx

Coding Text / Chat Image Input Agents Web Search

Context Window

1M Tokens

Latency (TTFT)

~0.6s Avg

Pricing (Input)

$0.30 / 1M

Pricing (Output)

$1.20 / 1M

M27H

MiniMax M2.7-highspeed

MiniMax / 200K ctx

Coding Text / Chat Agents

Context Window

200K Tokens

Latency (TTFT)

~0.3s Avg

Pricing (Input)

$0.60 / 1M

Pricing (Output)

$2.40 / 1M

M2.7

MiniMax M2.7

MiniMax / 200K ctx

Coding Text / Chat Agents

Context Window

200K Tokens

Latency (TTFT)

~0.5s Avg

Pricing (Input)

$0.30 / 1M

Pricing (Output)

$1.20 / 1M

Compare MiniMax

Z.ai (GLM) vs MiniMax

Compare side-by-side →

Highlights

🧠

1M Context

📊

Top SWE-bench

🎬

Multimodal

At a Glance

InterfaceWeb · API

Model SupportMiniMax M3, MiniMax M2.7-highspeed +1 more

PricePay-as-you-go · $22 · $54 · $130 / mo

Agent Mode

Terminal Access

IDE Integration

Multi-file Editing

Best ForFrontier AI platform with M3 — 1M context, coding & agents

Features

Models

MiniMax M31M context

MiniMax M2.7-highspeed200K context

MiniMax M2.7200K context

Capabilities

Code Generation

Agent Workflows

Image Input

Image Generation

Video Generation

Web Search

Audio / TTS

Key Features

1M token context window (M3)

MiniMax Code agent harness (desktop & web)

Agentic workflows & tool use

Native multimodal (text, image, video, audio)

Function calling

Streaming

Benchmarks

SWE-bench Verified74%

SWE-bench Multilingual72.5%

Terminal-Bench 2.047.9%

Real-World Task Ratings

Build Next.js SaaS

7/10

Fix Production Bug

7/10

Refactor Legacy Code

7/10

Create API

7/10

Write Tests

7/10

Cost Breakdown

API Input$0.30/1M tokens

Plus Plan$20/mo

Max Plan$50/mo

Est. Monthly Spend$20–120

Context & Memory

Context Window

1,000,000 tokens

Repo Understanding

★★★★★

Memory

★★★★★

Autonomy Score

Edit 1 File

8/10

Edit 10 Files

7/10

Create Feature

7/10

Deploy App

7/10

Fix Unknown Bug

7/10

Overall Average7.2/10

Community

3k+

GitHub Stars

Yes

Open Source

Platforms

Killer Features

Can Build Entire App

7/10

Vibe Coding Score

6/10

Full Breakdown

Basics

InterfaceWeb · API

CategoryAI Coding Assistants

Open source

Multi-language support

Websiteplatform.minimax.io

AI Capabilities

ModelsMiniMax M3, MiniMax M2.7-highspeed +1 more

Context window1,000,000 tokens

Agentic execution

Code generation

Multimodal input

Workflow

CLI / Terminal

IDE plugin

Web interface

API access

Desktop app

Pricing

APIFree

Go$22/mo

Explore$54/mo

Build$130/mo

Ideal For

API DevelopersAgent BuildersMultimodal AppsCost-Conscious Teams

Choose This Plan If…

1M token context window on the flagship M3 model
Strong coding and agentic benchmark performance
Native multimodal — text, image, video, and audio in one architecture
Competitive API pricing from $0.30/1M tokens
MiniMax Code agent harness bundles chat, memory, and scheduling in one place