LLM
VIDEO
SPEECH & MUSIC
Intelligence with everyone
M3 reaches frontier capability on coding and agentic tasks, introduces the brand-new MSA (MiniMax Sparse Attention) supporting up to 1M context, and is a natively multimodal model. It is the only domestic model combining all three Frontier essentials and will be the only open-source one in this class.
MiniMax
M3
MiniMax
M3
A frontier coding & agentic model built on a novel attention architecture (MSA) with 1M context
Frontier Coding / Agentic
Production-grade engineering, beyond code generation
MSA · 1M Context
Novel sparse attention that truly scales context
Multimodal
Step 0 joint training · 100T data · interleaved
Frontier Coding / Agentic
Production-grade engineering, beyond code generation
MSA · 1M Context
Novel sparse attention that truly scales context
Multimodal
Step 0 joint training · 100T data · interleaved
MiniMax
Code
MiniMax
Code
The coding harness built for MiniMax models
Build Agent Team
Evaluate tasks and assemble teams to solve problems.
Know Your Habits
Learn how you work and turn repetitive tasks into custom skills.
All in One Chat
Skills, memories, schedules, and more, all in one chat.
Build Agent Team
Evaluate tasks and assemble teams to solve problems.
Know Your Habits
Learn how you work and turn repetitive tasks into custom skills.
All in One Chat
Skills, memories, schedules, and more, all in one chat.
Token
Plan
Token
Plan
Revamped Token Plan · 2× Tokens · Frontier Models
Frontier Coding Capabilities
Frontier Model | MSA · 1M · Multimodal
Up to 1.7B tokens/month
Plus $20 ≈ 110K long documents
$20 = 10× Claude Pro
Same price, 10× the throughput
Frontier Coding Capabilities
Frontier Model | MSA · 1M · Multimodal
Up to 1.7B tokens/month
Plus $20 ≈ 110K long documents
$20 = 10× Claude Pro
Same price, 10× the throughput
Flagship Models
MiniMax's latest featured models — Language / Video / Speech & Music
Latest Research
Cutting-edge Tech · Model Capabilities · Engineering Practice
MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model
Read More
MiniMax Agent Team: Built for Long-Running Tasks and Continuous Evolution
Today we are introducing the overall upgrade of MiniMax Agent. We have given the upgraded Agent a new name: Mavis — MiniMax as a Jarvis, your AI butler.
Read More
Why Can't the MiniMax LLM Say "Ma Jiaqi"? Internal Investigation of Sparse Token Forgetting
The MiniMax M2 series has attracted widespread attention from the developer community. This article presents our internal investigation into the sparse token forgetting phenomenon.
Read More
MiniMax
A World-Leading General AI Technology Company
Founded in early 2022, MiniMax is driven by the mission to 'co-create intelligence with everyone,' dedicated to advancing the frontiers of AI and achieving Artificial General Intelligence (AGI). MiniMax has independently developed a series of multimodal foundation models with powerful code and Agent capabilities, as well as ultra-long context processing, capable of understanding, generating, and integrating multiple modalities including text, audio, image, video, and music.
Building on these proprietary models, MiniMax has launched a suite of AI-native products worldwide, including MiniMax Agent, Hailuo AI, MiniMax Audio, Talkie, and an open platform for enterprises and developers — delivering cutting-edge intelligent experiences to users around the globe.
0+
Countries & Regions Served
0M+
Global Individual Users
214,000+
Enterprise Clients & Developers
0+
Enterprise Coverage Countries