In November 2024, at the Comfy Community Summit Tokyo, an introduction to the MiniMax large model was presented. (Photo provided by MiniMax)
S4 Capital and MiniMax participated in a discussion on AI video capabilities at the Cannes Lions International Festival of Creativity in 2025. (Photo provided by MiniMax)
A bear performing somersaults, a lion leaping through flaming hoops, an acrobat walking a high wire - all seamlessly generated in a one-minute video. It's hard to believe this showcase of precision was created by three artists using Hailuo 02, MiniMax's video foundation model, capturing global attention upon its August 2024 debut.
Since then, the Chinese AI unicorn's Hailuo Video platform, through constant upgrading, has empowered creators worldwide to generate over 370 million videos. Hailuo AI Video ranked No. 1 in the global website traffic rating inJanuary 2025. In June 2025, the Hailuo 02 model secured the second place on Artificial Analysis' Video Model Arena - the industry's benchmark evaluation.
Renew Cost-Efficiency Record in Large Model Development Towards Inclusive AI for All
Adhering to the concept of "Intelligence with Everyone" (co-creating intelligence), MiniMax aims to help global creators fully unleash their imagination and enhance the quality of video creation, while reducing their entry barriers and cost burdens, making AI technology more accessible. This has driven researchers to continuously explore technological innovations, resulting in the development of an innovative model architecture called "Noise-aware Compute Redistribution (NCR)," which has led to astonishing improvements in the training and inference efficiency of Hailuo 02.
Thanks to architectural innovation, MiniMax's large model can generate native 1080p videos at a relatively affordable price. Three free versions of video generation are available on the Hailuo Video website, app, and open platform API: 768p-6 seconds, 768p-10 seconds, and 1080p-6 seconds.
According to a company executive of MiniMax, there are many cases that it is not MiniMax's technology that helps users, but it is often the global users who are actually assisting MiniMax. It is just because of their participation and feedback that there can be better intelligence. Only through the joint efforts of all parties can the vision of "co-creating intelligence" come true.
In the field of AI, many leading enterprises choose to consolidate their technological advantages through closed-source strategies, while MiniMax opts to open up the core capabilities of its large language model to the world. By lowering the threshold for technology access, MiniMax hopes to promote collaboration and innovation within the global developer community, and drive the entire industry towards more benign competition and development in terms of both technology and values.
To date, MiniMax's self-developed multimodal models and AI-native applications have provided services to over 157 million individual users from more than 200 countries and regions, as well as over 50,000 enterprise customers and developers from over 90 countries and regions.
"Open-sourcing high-level models enables the general public to experience the technical capabilities of excellent large models for free, thereby promoting the popularization and penetration rate of artificial intelligence technology," said the company executive of MiniMax. "At the same time, open-sourcing can also attract more developers to participate and jointly drive the rapid development of the large model industry," the executive added.
Experts from multiple countries discussed the interaction between humans and AI at a symposium held in November 2024. (Photo provided by MiniMax)
Go in Depth to Empower AI-oriented Global Transformation of Various Industries
MiniMax's multimodal large model and general-purpose Agent can understand and process various types of content such as text, images, videos, and music, and integrate them together to provide technical solutions based on the prompts and commands. It not only generates high-quality content but more importantly, it brings together the ability to handle multiple content forms, helping creative personnel to try different ideas more quickly, inspire creativity, and make the creation process smoother.
With its leading multimodal large model technology and AI solutions, MiniMax is deeply empowering the AI-oriented transformation of various industries worldwide. Recently, Monks under the world-known advertising and marketing company S4 Capital has entered into a strategic partnership with MiniMax, jointly committed to promoting the application and development of AI in the global media and advertising sectors. With MiniMax's advanced multimodal large model technology and universal Agent, the process that once took several days for proposal documents, conceptional works and initial draft generation has been shortened to mere seconds, enabling the design team to truly achieve the "light-speed" incubation of creativity.
Monks' creative team only needs to describe their ideas in a few simple sentences. Just a few seconds later, with MiniMax's multimodal large model technology, high-quality initial drafts of images, videos, or text can be generated accordingly. Meanwhile, MiniMax Agent can directly complete the production of a series of advertising materials and proposal documents, in forms of DOC and PPT files. This open and collaborative approach greatly simplifies the internal workflow of the team, allowing them to focus more on refining the creativity itself and saving a significant amount of time for clients.
"We are extremely delighted to enter into a strategic partnership with MiniMax. MiniMax's flagship video model and Hailuo AI offer our clients unparalleled AI- driven video generation quality. This collaboration not only enables us to experience their groundbreaking products in advance but also opens the door to the valuable expertise of the MiniMax team," said Rogier Bikker, Managing Director of Monks Greater China.
In the United States, the collaboration between Hedra, a renowned digital human company in Silicon Valley, and MiniMax is driving a profound transformation in the global content creation sector. A director with the company said that MiniMax Speech-02 performs particularly well when related to East Asian languages. This partnership not only integrates the world's advanced technological resources, but more strategically, helps build a diverse and resilient global AI technology supply chain. Through high-quality, multi-language AI voice generation technology, especially in emerging "voice-driven visual" applications such as virtual humans and the metaverse, it helps independent creators and small- and medium-sized teams to produce professional-level audio content at a lower cost, thereby better serving global users.
The technological capabilities of MiniMax are bringing about new changes in the global digital content creation field, heralding that content creation is about to enter a new stage full of infinite possibilities. (By Wu Shuang)