MiniMax Technology announced its development results since its establishment at the end of 2021, demonstrating its remarkable achievements in the field of artificial intelligence. They have built large models with trillions of parameters, covering text, speech and image fields, and launched multiple native applications and open platform API services to provide convenience for enterprises and developers. MiniMax's large models interact frequently with end users and process huge amounts of data, which reflects its technical strength and market influence. However, MiniMax founder IO also emphasized the direction and challenges of future development.
Since its establishment at the end of 2021, MiniMax has built trillion-parameter MoE text models, speech models, and image models, and launched native applications such as Hoshino and Conch AI while providing open platform API services for enterprises and developers.
Over the past 996 days, MiniMax's large models have had an average of 3 billion daily interactions with end users, processed more than 3 trillion text tokens, generated 20 million images and 70,000 hours of speech. These interactions demonstrate the company's rapid progress and broad impact in the field of artificial intelligence.
MiniMax founder IO emphasized that despite significant progress, the company is still a long way from achieving the goal of smart connectivity for 100% of the global population. To achieve this goal, MiniMax will continue to focus on technological innovation, especially in key areas such as reducing model error rates, supporting infinitely long input and output, and multi-modal interaction.
On this Partner Day, MiniMax released the video model abab-video-1, the music model abab-music-1, and updated the speech model abab-speech-1. In addition, MiniMax will release the multi-modal model abab7 using MoE + Linear Attention technology in the next few weeks.
In addition, MiniMax has also launched advanced speech and video models, including speech models that support multiple languages and emotional expressions, and video models with high compression rates and diverse styles. These models will be provided to users and developers through the company’s open platforms and products such as Hoshino and Conch AI.
MiniMax's rapid development in the field of artificial intelligence is remarkable, and its commitment to future technological innovation and pursuit of global intelligent connectivity goals are worth looking forward to. Continuous technological breakthroughs and product iterations will further consolidate its leading position in the industry.