MiniMax, a subsidiary of Shanghai Xiyu Technology Co., Ltd., launched the newly upgraded T2A-01 series of voice models globally on January 20, 2025, and launched Conch voice products. This series includes two models, T2A-01-HD and T2A-01-Turbo, which provide API services and users can choose according to their needs. As a general artificial intelligence technology company, MiniMax is committed to self-developing general large models of different modes. The speech model released this time supports 17 languages and hundreds of preset timbres, aiming to provide natural and smooth speech for corporate and individual users. Generate experiences.
On January 20, 2025, MiniMax, a subsidiary of Shanghai Xiyu Technology Co., Ltd., announced the simultaneous global launch of its newly upgraded T2A-01 series of voice models and the launch of Conch voice products. The T2A-01 series includes two models, T2A-01-HD and T2A-01-Turbo, for users to use. The API service has been simultaneously launched on the MiniMax open platform. Enterprises can choose based on their needs for audio quality and generation speed.
As a leading general artificial intelligence technology company, MiniMax focuses on self-developed general large models in different modes, including trillion-parameter MoE text large models, speech large models, and image large models. Based on these models, MiniMax has launched native applications such as Hoshino and Conch AI, and provides open platform API services for enterprises and developers. The T2A-01 series of speech models released this time not only have the characteristics of clear sound quality, natural rhythm, and accurate expression of emotions, but also support 17 languages including Chinese, Cantonese, English, Japanese, Korean, Arabic, and Spanish, and Hundreds of preset sounds provide business and individual users with a natural and smooth voice generation experience.
One of the highlights of Conch Voice is its powerful multi-lingual synthesis capability. With the support of the T2A-01 model, Conch Voice is ahead of similar products in terms of similarity, error rate and hearing evaluation. In multiple languages such as Chinese, Cantonese, English, Japanese, Korean and Arabic, the similarity and accuracy of Conch Voice are significantly ahead, and its comprehensive capabilities are comparable to the international leading model ElevenLabs. In addition, Conch Voice also has emotional understanding capabilities and can intelligently identify and reproduce subtle emotional differences in speech. Users can specify emotions as needed and generate speech output that accurately captures deep human emotions.
Conch Voice also provides users with rich tone selection and personalized adjustment functions. Users can filter according to language, accent, gender and age, select suitable timbres from more than 300 preset timbres, and fine-tune the timbre through effectors, such as adjusting the transparency and intensity of the timbre, and superimposing echo , broadcast, distortion, electronic music and other special effects to meet the needs of different scenes.
Conch Voice:
https://hailuoai.com/audio
Hailuo Audio (overseas version):
https://hailuo.ai/audio
Domestic API services:
https://platform.minimaxi.com/document/T2A%20V2
Overseas API services:
https://intl.minimaxi.com/document/T2A%20V2?key=66719005a427f0c8a5701643
Conch Voice launched by MiniMax provides users with an unprecedented voice generation experience with its powerful multi-language support, precise emotional expression and rich tone selection. Its leading technical strength and convenient API services are bound to set off a new wave in the field of speech synthesis.