Open-source music generator YuE2 tops SongBench, beating Suno v5 with symbolic planning
solyarisoftware · x · 2026-09-10
YuE2 is a new open-source music generation model that unifies symbolic and audio generation: it first writes an editable ABC symbolic score, then renders vocals and accompaniment, supporting creation, covering, and conversational agentic editing.
Key facts:
- 3.59B parameters, 28 layers; AR and NAR experts share attention with separate normalization
- On WildSongBench (192 prompts, 15 evaluated settings), YuE2 best-of-8 scores 6.9632 — the highest observed mean, beating Suno v5's 6.8721, as well as Mureka v9 and MiniMax 3
- Symbolic plans let users inspect and edit melody, rhythm, and chords before rendering
- Built by M·A·P, Tokenwave.AI, MBZUAI, and ACE STUDIO, with interactive demos and scores
More from Multimodal
- Video AI agent runs a physics engine, then paints your real room over the blocks — jonstephens85 · 2026-10-27
- Fish Audio ships open-weight Drama 3 TTS; creator shows turning old posts into narrated videos — alex_verem · 2026-10-27
- Redditor crafts photorealistic 'morning bus ride' short film entirely in ComfyUI — no3us · 2026-10-27
- New blur/anti-blur slider LoRA for MOSS lets you dial background sharpness with weights — NoMouse9610 · 2026-10-27
- PixGR: Pixel-space generative decoding for fast feed-forward 3D reconstruction — kwangmoo_yi · 2026-10-27
- Krea-2 Turbo + ComfyUI for product ads: prompts and results shared — comfy-king-999 · 2026-10-27