한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
as a small company focusing on ai technology research and development, minimax's first products "video-01" and "music-01" will open a new chapter in the era of multimodal models. the video model "video-01" can generate five-second videos with its native high-resolution and high-frame rate video generation capabilities. users can experience the product through the minimax official website. the music model "music-01" focuses on end-to-end multifunctional music generation, providing users with a new creative experience.
although minimax has only released the first version of the product, its ambition and technical strength for future development are exciting. yan junjie, founder and ceo of minimax, said, "what we are showing to the public is only the first version of the product, and updated versions will be gradually released in the future." he also emphasized that future commercialization will mainly be divided into two forms: based on the company's open platform and the accumulation of more than 2,000 customer partners, many of which are well-known corporate users who are also willing to use voice recognition capabilities, and the other is to introduce advertising mechanisms in its own products.
as a leader in ai technology, minimax is actively exploring the application areas of multimodal models. "this is just the beginning," yan junjie said, "in the future, we will continue to work hard to improve model performance and release more corresponding products."
minimax's user interaction data shows that the model handles more than 3 billion customer interactions. this means that minimax users have experienced the convenience of multimodal interaction. however, user penetration and usage depth need to be further improved. "it is not difficult to find in life that text interaction is only a small part, and more is voice and video interaction. multimodal content, such as sound, graphics and video, has become the mainstream of information transmission," yan jün jie explained, which is why minimax will continue to explore the direction of multimodal applications.
in order to improve model performance, minimax continuously optimizes technology and sets new goals. "fast" is the core technology research and development goal of minimax's underlying large model. in the past few years, minimax has made certain breakthroughs in technology. the application of technologies such as moe (mixed expert architecture) and linear attention has laid a solid foundation for future technological development. the new generation of models developed by minimax based on moe+ linear attention are considered to be comparable to the level of gpt-4o.
minimax has completed three rounds of financing in the past three years, with investors including well-known companies such as tencent and mihoyo. today, minimax is valued at more than $2.5 billion.
with the continuous development and application of ai technology, multimodal models will become the core direction of future technological development. minimax will bring users a more convenient experience through continuous exploration and breakthroughs.
