MOSS-SoundEffect-v2.0: Text-to-Sound Effect Generation

MOSS-SoundEffect-v2.0 has been released, utilizing a Diffusion Transformer architecture to create high-quality sound effects up to 30 seconds long.

Compiled by Sergey KostenchukPublished 2026-07-14Updated 2026-07-14

2026-07-14 Multimodal HuggingFace

🎨 MOSS-SoundEffect-v2.0: Text-to-Sound Effect Generation

MOSS-SoundEffect-v2.0 has been released — a new model for creating SFX from text descriptions. By transitioning to a Diffusion Transformer (DiT) architecture with Flow Matching, the model ensures high realism and stability of sound scenes (48 kHz, up to 30 seconds).

🌍 The transition to DiT brings SFX generation quality closer to professional sound engineering standards.

👤 Create complex sound environments using simple text via Python or ComfyUI.

Source 1: https://github.com/OpenMOSS/MOSS-TTS/tree/main/moss_soundeffect_v2 Source 2: https://huggingface.co/OpenMOSS-Team/MOSS-SoundEffect-v2.0

Sources