Kimi-Audio Technical Report

Benchmark Model Rank Results
audio-understanding-reasoning-on-mmauKimi-Audio (8.2B)Accuracy (Avg, Test-mini): 68.20
speech-recognition-on-librispeech-test-cleanKimi-Audio#18Word Error Rate (WER): 1.28
speech-recognition-on-librispeech-test-otherKimi-Audio#7Word Error Rate (WER): 2.42