Fish Audio

Headlines · 3

1. 6d
   INT21 says AI agents built a Drama 3 Rust runtime in one hour

   INT21 says its SwarmOS orchestrated AI agents that autonomously wrote, tested and optimized a Rust inference runtime for Fish Audio’s Drama 3 in about an hour. Its workflow generates multiple candidate implementations, tests them on hardware including NVIDIA B200 GPUs, retains working versions and discards failed ones. INT21’s benchmark reported time to first audio of 15.11–23.58 milliseconds, compared with 37.79–38.61 milliseconds for SGLang-Omni. Each test used one speech stream, a one-frame first chunk and 20 warm trials per prompt across three prompts; the results are INT21’s own reported figures.

   Wccftech · 🔥 0
2. Oct 13
   AI editing tools compete as spending shifts from editors to compute

   A Dingjiao One industry analysis describes a test in which an editor had Claude Opus 5.5 produce a short video end to end: the model called Fish Audio to synthesize voiceover and HeyGen to drive a digital avatar, then analyzed two reference videos frame by frame, turned the patterns into code for motion graphics and sound effects, and packaged the method as a reusable Skill. It notes that in September 2026 CapCut (Jianying) launched an AI creation workspace called Hub, an AI editing assistant and an AI Ultra membership (199 yuan per month with 22,100 AI credits), that Adobe extended Creative Agent to Premiere and other products in June, and that DaVinci Resolve 21.1, released in September, integrates external agents such as Claude, Claude Code and Codex. The piece groups competitors into platform ecosystems, professional software, AI-native startups and upstream model providers, and cites vendor demos and customer cases showing shorter editing hours, while cautioning that such figures do not apply to every project.

   虎嗅 AI · 🔥 0
3. Oct 12
   Fish Audio releases open-weight Drama 3 speech model with 35B MoE backbone

   Fish Audio has published a technical report for Fish Audio Drama 3, an open-source speech generation model that it says turns text into directed performances. The model uses a 35B-parameter mixture-of-experts backbone activating 3B parameters per token, clones voices from multiple references, follows natural-language direction, and adds multi-turn revision, multi-character scene dubbing, editing by infilling and streaming text input. The report pairs a hybrid linear-attention backbone with a new codebook multi-token-prediction audio decoder, and describes training via three-stage pre-training, instruction fine-tuning and large-batch reinforcement learning rewarded by a single learned human-preference model. Weights are said to be available on GitHub and Hugging Face with inference through SGLang-Omni, and the authors report 271 frames per second for single-stream decoding on one H200 versus 177 for Fish Audio S2.

   Hugging Face · Papers · 🔥 0

Experience and discussion from the community

Share my Fish Audio experienceAsk about Fish Audio