{"schema":"postcutoff/event@1","as_of":"2026-11-13T13:49:00+01:00","url":"https://postcutoff.com/e/2026-09-09-yue2-open-music-model/","md":"https://postcutoff.com/e/2026-09-09-yue2-open-music-model/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-09-09-yue2-open-music-model","date":"2026-09-09","date_precision":"day","short_title":"YuE2: open-weights song model that plans an editable score first, claims top WildSongBench score over Suno v5","deck":null,"takeaway":null,"category":"open-source","category_label":"Open source","importance":3,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":[{"n":1,"title":"GitHub: multimodal-art-projection/YuE (YuE2)","url":"https://github.com/multimodal-art-projection/YuE","type":"code","group":"primary","domain":"github.com"},{"n":2,"title":"Hugging Face: m-a-p/YuE2-3B","url":"https://huggingface.co/m-a-p/YuE2-3B","type":"code","group":"primary","domain":"huggingface.co"},{"n":3,"title":"Demo page","url":"https://map-yue2.github.io","type":"official","group":"primary","domain":"map-yue2.github.io"},{"n":4,"title":"YuE (v1) paper, arXiv 2503.08638","url":"https://arxiv.org/abs/2503.08638","type":"paper","group":"primary","domain":"arxiv.org"}],"official":4,"filed":"2026-09-29","updated":"2026-09-29","orgs":["Multimodal Art Projection (M-A-P)","HKUST"],"title":"YuE2: open-weights song model that plans an editable score first, claims top WildSongBench score over Suno v5","summary":"The M-A-P research community (HKUST and partners) released YuE2, a ~3-4B open-weights song generator that first writes an editable melody-and-chord score (ABC notation) and then renders full songs with vocals and accompaniment at 48 kHz stereo, with zero-shot covers and conversational \"agentic\" music editing; its authors report it beat all evaluated open and proprietary systems, incl. Suno v5, on their 192-prompt WildSongBench (best-of-8).","key_facts":["Weights published on Hugging Face (m-a-p/YuE2-3B, YuE2-Vae) around 2026-09-09; tech report 2026-09-26, arXiv 2609.33757 on 2026-09-29","Architecture: AR-NAR Mixture-of-Transformers generating symbolic scores and acoustic latents via flow matching; model card lists ~4B parameters despite the '3B' name","Self-reported WildSongBench (192 prompts, run 2026-09-12): YuE2 best-of-8 SongBench avg 6.9632 vs Suno v5 6.8721","Zero-shot covers: 0.647 CLEWS mAP on 948 works (self-reported)","Lyrics in English and Mandarin; instrumental generation added 2026-09-25; companion MERT-v2 and SheetSage2 (audio-to-score) models","License: weights CC BY-NC 4.0 (commercial license available; README says outputs may be monetized royalty-free), code Apache 2.0","Community ports within days: GGUF, MLX, ComfyUI, many genre LoRAs"],"key_numbers":[],"tags":["music-generation","open-weights","audio","symbolic-music","agents"],"science":null,"body_md":"## What happened\n\nYuE (Jan 2025) was the first open lyrics-to-full-song model. YuE2 changes the approach: it writes a symbolic plan (melody and chords) that users can edit, then renders audio from it, which enables covers, score edits and chat-driven revisions. Benchmark claims are the authors' own and have not been independently reproduced. The arXiv id 2609.33757 is taken from the GitHub README and was not opened.\n\n## Why it matters\n\nIt is the strongest claim yet that an open model runnable on one consumer GPU matches the leading commercial song generator, released the same day Suno moved to licensed-data v6. Its non-commercial weight license limits commercial use.","disputed":[],"related":[{"id":"2026-09-09-suno-v6-licensed-music-model","url":"https://postcutoff.com/e/2026-09-09-suno-v6-licensed-music-model/","date":"2026-09-09","date_precision":"day","short_title":"Suno launches v6, its first music models trained on licensed music","deck":null,"takeaway":"v6 is the clearest test yet of a licensed-training business model for generative media; the new Sony/UMG suit tests whether \"clean-room\" retraining on licensed data (but with learnings from older models) is enough.","category":"media-generation","category_label":"Media generation","importance":3,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":5,"official":0,"filed":"2026-09-29","updated":"2026-09-29","orgs":["Suno","Warner Music Group","BMG","Believe"]},{"id":"2026-01-28-ace-step-1-5","url":"https://postcutoff.com/e/2026-01-28-ace-step-1-5/","date":"2026-01-28","date_precision":"day","short_title":"ACE-Step 1.5: MIT-licensed song generator that runs on consumer GPUs","deck":null,"takeaway":"It made near-commercial song generation practical on laptops and gaming GPUs under a permissive license, the open alternative to Suno/Udio in the year the commercial services moved to licensed data.","category":"open-source","category_label":"Open source","importance":3,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":4,"official":4,"filed":"2026-09-29","updated":"2026-09-29","orgs":["ACE Studio","StepFun"]}],"people":[],"posts":[],"videos":[],"models":[{"id":"yue2","name":"YuE2-3B","url":"https://postcutoff.com/m/yue2/"}],"changes":[{"date":"2026-09-29","type":"filed","text":"Created"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-09-29","run":null,"sources_read":null,"updated":"2026-09-29","human_review":null,"version":{"date":"2026-10-04"}},"gaps":[{"model_id":"gpt-6-astra","name":"GPT-6 Astra","cutoff":"2026-04","days_after":132,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":71,"in_training_data":false},{"model_id":"gemini-4-hydrogen","name":"Gemini 4 Hydrogen","cutoff":"2026-03","days_after":162,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":101,"in_training_data":false}],"short_url":null}