Open-source YuE2 generates full songs via symbolic planning, taking on Suno V5
bdsqlsz · x · 2026-09-10
The multimodal-art-projection team released YuE2, an open-source music generation model claiming frontier quality. Give it lyrics and a style prompt: it first writes a symbolic melody-and-chord plan, then renders a complete song with vocals and accompaniment. New features include zero-shot covers, agentic music editing, an agent skill, and companion models (MERT2, SheetSage2), benchmarked competitively against proprietary systems on WildSongBench — with the author openly asking whether it beats Suno V5.
More from Multimodal
- Cartesia explains why benchmarking TTS is extremely hard: no single number captures voice quality — saranormous · 2026-09-16
- YuE2 music sampling only hits ~7 tokens/s on RX 9070 despite mostly idle VRAM — Prestigious-Kick7291 · 2026-09-16
- Pixio launches AE plugin: chat agent builds native layers and keyframes in your timeline — tsi_org · 2026-09-16
- Midjourney style ref + Gemini photorealism: a two-step image workflow — michaelrabone · 2026-09-16
- Seedance 2.5 recreates a GTA-style stealth mission with flawless phone tracking — SimplyAnnisa · 2026-09-16
- Nex-N2.5 Pro builds an interactive 3D mechanical garden by iterating on its own output — nikola_mr64990 · 2026-09-16