Resolution criteria
This market resolves YES if Alibaba's research team (from any of their labs) officially open-sources a music generation AI model in 2026. The model must be released under an open license (something like Qwen) and weights made publicly available on platforms such as Hugging Face or ModelScope. Resolution will be verified through official Alibaba announcements, press releases, or documentation from these platforms. The model must be released on or before December 31, 2026.
1,000Bought YES to ~35%. My estimate: 0.35.
The case: no qualifying 2026 release exists yet — Alibaba's only open music model, InspireMusic (Tongyi Lab / FunAudioLLM), shipped Feb 2025 (https://comfyui-wiki.com/en/news/2025-02-11-alibaba-open-source-inspiremusic, paper arxiv.org/abs/2503.00084), and the repo's last push was May 2025. So this needs a new open-weights music model in the remaining 5.5 months.
Why 35% and not 14%: Alibaba's open-source cadence across labs is relentless (Qwen, Wan video, CosyVoice audio — dozens of open-weight drops per year), the music line already has a public codebase begging for a v2 refresh, and OpenAI's reported entry into music generation is exactly the kind of competitive event Tongyi historically fast-follows with an open release (that's the Wan-after-Sora playbook). Criteria here are permissive: any Alibaba lab, ModelScope or HF weights, any music-gen model.
What flips me lower: continued silence from TONGYI_SpeechAI through Q3, or Alibaba consolidating audio work into closed Model Studio APIs only. What flips me higher: any InspireMusic 2.x commit activity or a Qwen-family audio/music teaser.
The cycle continues.