Skip to main content
M

MiniMax

MiniMax delivers a high‑capacity coding and agentic AI model that leverages a novel multi‑scale attention (MSA) architecture to handle up to 1 million tokens of context. The platform enables developers to generate, understand, and refactor code across extensive codebases without losing context, streamlining large‑scale software development workflows.

Singapore, SingaporeFounded 202227230K+ followers
Updated 1 month ago

Funding

Funding not disclosed

Funding rounds are not available yet.

Founders

Product

Problem

Developers and enterprises often face limitations with existing AI models that cannot handle extremely long codebases or complex multimodal tasks, leading to fragmented workflows and reduced productivity.

Solution

MiniMax offers a suite of frontier AI foundation models, highlighted by the M3 model, which combines a novel Multi‑Scale Attention (MSA) architecture with up to 1 million token context windows. This ultra‑long context capability enables comprehensive code generation, refactoring, and reasoning over large projects without truncation. The models are natively multimodal, supporting text, audio, image, video, and music, allowing developers to integrate diverse data types within a single AI pipeline. MiniMax provides an open API platform and specialized products such as MiniMax Code, Hub, Audio, and Talkie, delivering flexible access for both individual developers and large enterprises. Pricing is available via pay‑as‑you‑go token billing or subscription plans, letting users scale usage according to their needs.

Target Audience

Primary customers are software developers, AI product teams, and enterprises that require large‑scale code assistance, multimodal generation, or autonomous AI agents within their development pipelines.

Features

  • M3 model with 1 M token context window for extensive code and reasoning tasks
  • Multi‑Scale Attention (MSA) architecture delivering high‑capacity, low‑latency inference
  • Native multimodal support (text, audio, image, video, music) in a single model
  • Agentic capabilities for long‑running, autonomous workflows (e.g., Mavis AI butler)
  • Comprehensive API with language, video, speech, and music endpoints
  • Specialized product suites (MiniMax Code, Hub, Audio, Talkie) for targeted developer experiences
  • Flexible pricing: real‑time per‑call token billing and monthly subscription quotas
This profile is AI-generated and may contain inaccuracies.