On April 26, Kimi announced its new open-source project—a universal audio foundation model, Kimi-Audio. The model supports various tasks, including speech recognition, audio comprehension, audio-to-text conversion, and voice interaction.

Editor: Gao Han