MiniMax(海螺AI)

3dys agoupdate 2,622 00

Building AGI with our mission Intelligence with Everyone. Global leader in multi-modal models and AI-native products with over 150 million users.

Location:
China
Language:
en,zh
Collection time:
2026-01-06
Other sites:
MiniMax(海螺AI)MiniMax(海螺AI)

MiniMax AI is a prominent Chinese artificial intelligence company with a focus on developing advanced “full-stack” multimodal large models. Officially known as Shanghai Xiyu Technology, it has quickly become a significant player in the global AI landscape by developing foundational models, AI-native consumer products, and enterprise services.

For a quick overview, here are the key facts:

AspectDetails
Full NameShanghai Xiyu Technology Co., Ltd. (稀宇科技)
Founded2021/2022
Core Philosophy“Good technology is the product”
Key ModelsText: MiniMax M2; Video: Hailuo-02; Audio: Speech-02
Key ProductsMiniMax Chat, Hailuo AI, Talkie/Starfield, Enterprise APIs
Global ReachUsers in 200+ regions; 70%+ revenue from overseas markets
Commercial StatusIPO process for Hong Kong Stock Exchange completed in late 2025

 

🏢 Company Background and Strategy

MiniMax is described as a “born-global” AI company dedicated to creating internationally competitive general-purpose models from its inception. A defining characteristic is its status as a “Pureplay” AI model company, meaning its core resources and business model are intensely focused on foundational AI research rather than diversifying into other internet services. This focus is supported by a young (average age 29), research-heavy team where about 74% of its 385 employees are in R&D. The company adheres to a “good technology is the product” philosophy, driving rapid iteration and flat, efficient organizational management.

🧠 Core Technology and Models

The company’s strength lies in its self-developed trio of powerful models covering all major AI modalities, forming the foundation for its products.

  1. MiniMax M2 (Text Model): This flagship large language model (LLM) was released and open-sourced in October 2025. Upon release, it ranked among the top five globally and was #1 among open-source models on the Artificial Analysis benchmark. It excels in programming applications, ranking third globally for token usage in programming scenarios on the OpenRouter platform.

  2. Hailuo-02 (Video Model): Launched in June 2025, this advanced video generation model quickly achieved a global rank of #2 on the Artificial Analysis video evaluation benchmark. It is noted for its understanding of complex physical motion and光影 effects, and has already generated over 590 million videos.

  3. Speech-02 (Audio Model): An advanced audio model built on a Transformer architecture. Its predecessor and current version have collectively generated over 2.2 billion hours of synthetic speech, powering highly realistic audio applications.

📱 Product Ecosystem

MiniMax builds a complete commercial ecosystem on top of its core models, serving both consumers (C端) and businesses (B端).

  • Consumer Products: These include the MiniMax Chatbot for general conversations and tasks, Hailuo AI for image and video generation, and interactive social apps like Talkie and Starfield.

  • Enterprise & Developer Services: Through its Open Platform, MiniMax provides API access to its full suite of models. It serves over 130,000 enterprise clients globally. The platform distinguishes between domestic (api.minimaxi.com) and international (api.minimax.io) API endpoints.

  • AI Agent Innovation: The company has introduced a sophisticated MiniMax Agent, designed as a “super employee” or “executable digital worker.” It can handle long-term, complex projects like developing a complete concert seat selection system with payment functions in just 30 minutes.

📊 Commercial Performance and Market Position

Financially, MiniMax has demonstrated explosive growth:

  • User Base: It has amassed over 212 million registered users and serves clients in more than 200 countries and regions.

  • Revenue: Revenue grew from $3.46 million in 2023 to $30.52 million in 2024 (a 782% increase). For the first nine months of 2025, revenue reached $53.44 million, growing 175% year-over-year.

  • Globalization Success: A key indicator of its global strategy is that over 70% of its revenue comes from overseas markets.

  • Efficiency: Despite rapid growth, the company has controlled costs effectively. In 2025, sales and marketing expenses decreased by 26% even as revenue surged over 170%, indicating strong product-led growth.

  • Investor Confidence: The company is backed by major institutions like Alibaba, Tencent, and Hillhouse Capital and has passed the listing hearing for the Hong Kong Stock Exchange.

In summary, MiniMax AI stands out as a full-stack, multimodal AI pioneer with robust in-house technology, a rapidly growing global product ecosystem, and a clear path as a publicly-traded “Pureplay” AI contender on the world stage.

data statistics

Relevant Navigation

Doubao(豆包)

Doubao(豆包)

Doubao is an AI tool developed by ByteDance based on its Skylark model. It provides functionalities such as a chatbot, a writing assistant, and an English learning assistant. It can answer various questions and engage in conversations to help users obtain information. The tool is accessible via a web platform on browsers, dedicated client applications for Windows and macOS computers, as well as iOS and Android mobile platforms. In 2016, ByteDance established its AI Lab, focusing on research in natural language processing, machine learning, data mining, and related fields.
DeepSeek

DeepSeek

DeepSeek was founded in 2023 with a focus on researching world-leading fundamental models and technologies for artificial general intelligence (AGI), tackling cutting-edge challenges in AI. Leveraging self-developed training frameworks, proprietary AI computing clusters, and tens of thousands of computing units, the DeepSeek team released and open-sourced multiple billion-parameter large models within just half a year. These include the general-purpose large language model DeepSeek-LLM and the code-focused model DeepSeek-Coder. In January 2024, DeepSeek also pioneered the open-sourcing of China's first Mixture of Experts (MoE) large model, DeepSeek-MoE. All these models have demonstrated outstanding performance, surpassing comparable models in both public benchmark evaluations and real-world generalization tasks beyond their training data. Chat with DeepSeek AI and easily access its API.

No comments

none
No comments...