Kimi-Audio

model paper
Moonshot AI
Universal audio foundation model for understanding, generation, and conversation. Trained on 13M+ hours of audio data.
Paper (arXiv)GitHubHuggingFace

Outputs 2

Kimi-Audio-7B-Instruct

model
Parameters 7B

Kimi-Audio Technical Report

paper

arXiv: 2504.18425

audioopen-weight