Back to news
launchiFlytek2026-11-12

iFlytek Puts SparkAudio 1.0 Speech Foundation Model on API with 99 Languages

iFlytek announced on November 12 that SparkAudio 1.0, a speech foundation model supporting 99 languages and 202 dialects, is now available via API on its open platform for transcription, translation, diarization and audio question answering.

iFlytek announced on November 12 that SparkAudio 1.0, its speech foundation model, is available via API on the open platform. It accepts text and speech input and returns text, supporting 99 languages and 202 dialects, and was trained entirely on domestic compute.

讯飞星火语音大模型API