Xiaohongshu Open-Sources FireRedAudio, a 9B Unified Audio Language Model
aigclink · x · 2026-08-31
Xiaohongshu's FireRedTeam has open-sourced FireRedAudio, a general-purpose audio language model built on a shared 9B-parameter LLM backbone: an Audio Encoder handles understanding while a RedAE pathway handles generation, with the two representations decoupled yet sharing the same reasoning core.
A single model supports ASR, audio understanding, zero-shot TTS, instruct TTS, semantic/acoustic speech editing, and accurate temporal grounding over recordings up to one hour. In practice, a one-hour podcast can be transcribed into a timestamped professional show script, with voice cloning, text-designed new voices, and audio editing — output quality approaches manual interview-material editing.
Related event: Xiaohongshu Open-Sources FireRedAudio, a 9B Unified Audio-Language Model(2 posts)→
More from Multimodal
- Orchestrating a 100-image portrait of humanity for aliens with ChatGPT, Argon and p5.brush — dx8xb · 2026-10-16
- Indie Creator Builds an AI-Native Animated Show Solo — ComprehensiveLime654 · 2026-10-16
- Ideogram teases 4.5 arrival and open weights: 'Not yet, but soon' — East_Shoe8814 · 2026-10-16
- Opus 5.5 turns heads with surprisingly polished vertical video demo — FinanceYF5 · 2026-10-16
- 'Celestial Ocean Reverie' prompt template blends starry skies with dreamlike seascapes — LudovicCreator · 2026-10-16
- 'Thicc Lil' Red': an AI-made musical meme reimagining Little Red Riding Hood — bennash · 2026-10-16