MiniMax 发布开源音乐模型 Music3，可一次生成五分钟歌曲
MiniMax Releases MiniMax-Music3: An Open-Weights Music Model Generating Complete Five-Minute Songs From Lyrics and a Structured Caption
做音乐生成或音频 Agent 的同学值得关注：开源权重、一次生成五分钟完整歌曲，还给了三条可落地的部署路径，直接拿去跑一版试试。
MiniMax released MiniMax-Music3, an open-weights text-to-music model. The model takes two separate inputs: lyrics carrying section tags, and a detailed music description. It returns a complete song of up to five minutes in a single generation, as 32 kHz, 16-bit stereo WAV. The architecture pairs a Hybrid-LM, an 8B Global LLM with a 0.6B Local LLM, with a continuous synthesis stack built on flow matching and a Flow-VAE. Weights, inference code and three documented serving paths shipped the same day.
MiniMax 发布了 MiniMax-Music3，一个开放权重的文本转音乐模型。该模型接受两个独立输入：带有段落标签的歌词，以及详细的音乐描述。它能在单次生成中返回一首最长五分钟的完整歌曲，格式为 32 kHz、16 位立体声 WAV。其架构将混合语言模型（一个 8B 全局 LLM 和一个 0.6B 局部 LLM）与基于流匹配和 Flow-VAE 的连续合成栈配对。权重、推理代码和三条文档化的服务路径同日发布。
Is it deployable?
它可部署吗？
Yes, MiniMax published usable weights, inference code and three documented serving paths on day one, so this is deployable now rather than a research preview.
是的，MiniMax 在发布当天就提供了可用的权重、推理代码和三条文档化的服务路径，因此这现在就可以部署，而不是研究预览。
- Company level: Solo creators, indie studios and mid-market teams can ship on it directly. The MiniMax-Music3 Community License permits commercial use, but it requires you to display ‘MiniMax-Music3’ prominently in the product UI, and any organization whose aggregate yearly revenue from those products exceeds US$ 20 million must obtain separate prior written authorization from MiniMax. Anyone hosting third-party generation must also implement and maintain safeguards against infringing outputs.
- Industries: Game development, advertising and brand agencies, short-form video and creator tools, e-learning, podcasting, fitness and wellness apps, retail in-store audio, and music-tech SaaS.
- Applications: Background scoring for UGC video, adaptive game and level music, localized ad beds and sonic branding, scratch and demo tracks for songwriters, mood-conditioned playlist generation, and offline batch generation where per-song API cost is the constraint.
- 公司层面：独立创作者、独立工作室和中型团队可以直接基于它发布产品。MiniMax-Music3 社区许可证允许商业使用，但要求在产品 UI 中显著展示“MiniMax-Music3”，并且任何组织如果其来自这些产品的年度总收入超过 2000 万美元，必须事先获得 MiniMax 的单独书面授权。任何托管第三方生成服务的机构还必须实施并维护防止侵权输出的保障措施。
- 行业：游戏开发、广告和品牌代理、短视频和创作者工具、电子学习、播客、健身和健康应用、零售店内音频，以及音乐科技 SaaS。
- 应用：UGC 视频的背景配乐、自适应游戏和关卡音乐、本地化广告位和声音品牌、词曲作者的草稿和演示曲目、基于情绪条件的播放列表生成，以及按歌曲 API 成本受限的离线批量生成。
The Architecture
架构
MiniMax-Music3 combines a hierarchical autoregressive stack with a continuous synthesis path.
MiniMax-Music3 结合了分层自回归堆栈和连续合成路径。
The training tokenizer uses eight layers of residual vector quantization (RVQ). The first, semantic codebook has 16,384 entries and carries core musical semantics and structure. The remaining seven acoustic codebooks have 1,024 entries each and encode residual detail. Training optimizes the semantic layer first, then all eight jointly.
训练分词器使用八层残差向量量化（RVQ）。第一层语义码本有 16,384 个条目，承载核心音乐语义和结构。其余七个声学码本各有 1,024 个条目，编码残差细节。训练首先优化语义层，然后联合优化全部八层。
The Hybrid-LM splits the modeling problem. An 8B Global LLM predicts the first RVQ codebook frame by frame and holds long-range structure; a 0.6B Local LLM predicts the remaining codebooks within each frame. The model card and license state the Global LLM is initialized from Qwen3-8B; the MiniMax Research post says Qwen3.5-8B, so treat the exact base checkpoint as unsettled.
混合语言模型将建模问题拆分。一个 8B 全局 LLM 逐帧预测第一个 RVQ 码本并保持长程结构；一个 0.6B 局部 LLM 预测每帧内的其余码本。模型卡和许可证声明全局 LLM 从 Qwen3-8B 初始化；MiniMax 研究博客则称 Qwen3.5-8B，因此具体的基础检查点尚不确定。
The synthesis stage is the more interesting design choice. Rather than decoding from discrete RVQ tokens, MiniMax fuses the final hidden states of both LLMs and conditions a 2.4B flow-matching module on them, which maps into a latent space decoded by a 123M Flow-VAE inherited from MiniMax Speech. At inference the discrete tokenizer decoder is not loaded at all.
合成阶段是更有趣的设计选择。MiniMax 并非从离散的 RVQ 令牌解码，而是融合两个 LLM 的最终隐藏状态，并在此基础上调节一个 2.4B 的流匹配模块，该模块映射到由继承自 MiniMax Speech 的 123M Flow-VAE 解码的潜在空间。在推理时，离散分词器解码器完全不被加载。
Two-input control
双输入控制
Lyrics carry the words and section tags on their own lines: [Intro], [Verse], [Pre-Chorus], [Chorus], [Post-Chorus], [Bridge], [Instrumental], [Solo], [Outro]. A separate Structured Caption carries Global Metadata, Vocal Details and Arrangement. MiniMax also ships a music-caption-rewriter agent skill that expands a short description into that three-part format offline.
歌词在各自的行上携带文字和段落标签：[前奏]、[主歌]、[预副歌]、[副歌]、[后副歌]、[桥段]、[器乐]、[独奏]、[尾声]。一个单独的结构化字幕携带全局元数据、人声细节和编曲。MiniMax 还附带了一个音乐字幕改写代理技能，可离线将简短描述扩展为三部分格式。
Interactive explainer
交互式说明
Running it
运行它
Three documented paths. SGLang-Omni is the reference server; the GitHub page specifies two CUDA GPUs, with GPU 0 running Qwen3 and RVQ autoregressive generation and GPU 1 running flow matching and DAV decoding. The diffusers modular pipeline fits under 24 GB VRAM at full precision, about 22 GB with automatic CPU offload, and down to 8 GB with leaf-level group offloading. ComfyUI has a native Text to Music template using repacked FP16/INT8 weights from Comfy-Org.
三种有记录的路径。SGLang-Omni 是参考服务器；GitHub 页面指定了两个 CUDA GPU，GPU 0 运行 Qwen3 和 RVQ 自回归生成，GPU 1 运行流匹配和 DAV 解码。diffusers 模块化流水线在完整精度下适配 24 GB 以下显存，启用自动 CPU 卸载时约 22 GB，启用叶级分组卸载时可降至 8 GB。ComfyUI 有一个原生的文本到音乐模板，使用来自 Comfy-Org 的重新打包的 FP16/INT8 权重。
Key Takeaways
关键要点
- Open-weights model generating complete five-minute songs at 32 kHz, 16-bit stereo, released August 13, 2026.
- Hybrid-LM design: 8B Global LLM plus 0.6B Local LLM, feeding 2.4B flow matching and a 123M Flow-VAE.
- Synthesis runs on fused continuous hidden states, skipping the discrete tokenizer decoder entirely.
- Runs on two GPUs via SGLang-Omni, under 24 GB via diffusers, or 8 GB with group offloading.
- Commercial use allowed with visible attribution; above USD 20M revenue needs written authorization.
- 开放权重模型，生成完整的五分钟歌曲，32 kHz、16 位立体声，于 2026 年 8 月 13 日发布。
- 混合 LLM 设计：8B 全局 LLM 加 0.6B 局部 LLM，馈入 2.4B 流匹配和 123M Flow-VAE。
- 合成基于融合的连续隐藏状态运行，完全跳过离散分词器解码器。
- 通过 SGLang-Omni 在双 GPU 上运行，通过 diffusers 在 24 GB 以下运行，或通过分组卸载在 8 GB 下运行。
- 允许商业使用，需注明出处；收入超过 2000 万美元需书面授权。
Check out the Model on HF and GitHub Repo. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
在 HF 和 GitHub 仓库上查看模型。此外，欢迎在 Twitter 上关注我们，别忘了加入我们的 15 万+ ML SubReddit 并订阅我们的新闻通讯。等等！你在 Telegram 上吗？现在你也可以在 Telegram 上加入我们。
Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us
需要与我们合作推广你的 GitHub 仓库、Hugging Face 页面、产品发布或网络研讨会等吗？请联系我们。
The post MiniMax Releases MiniMax-Music3: An Open-Weights Music Model Generating Complete Five-Minute Songs From Lyrics and a Structured Caption appeared first on MarkTechPost.
文章《MiniMax 发布 MiniMax-Music3：一个开放权重音乐模型，从歌词和结构化字幕生成完整的五分钟歌曲》最初出现在 MarkTechPost 上。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力