Meta 发布开源权重模型 Muse Spark Open，对标 Qwen、GLM 与 Kimi
Muse Spark Open is Meta's answer to Qwen, GLM, and Kimi in the open-weight race
- 1Meta 发布开源权重模型 Muse Spark Open，816B 参数（42B 激活），主打编程与智能体任务
- 2官方称其在智能体、知识与推理、多模态类别领先 Qwen 3.8、GLM 5.3 和 Kimi K3，编程基准落后部分对手
- 3官方口径同时承认，该模型性能明显不及 Meta 付费的 Muse Wonder Full 系列
- 4权重上架 Hugging Face，免费使用须标注出处，年收入超 2000 万美元的 API 提供方需另与 Meta 签协议
Meta 重回开源权重赛道对抗 Qwen、GLM、Kimi 阵营，同时用许可证条款为大规模商用保留收费空间。
- 影响：开源权重让企业可自行部署多模态与百万 token 上下文模型，但本地满配需要两台各 512 GB 统一内存的 Mac Studio；许可证的标注义务和 2000 万美元收入门槛会进入商用合规考量。
- 参考：可考虑先核对许可证的标注义务与 2000 万美元收入门槛是否触及自身业务，再挑智能体或多模态场景做小规模试点；本地部署前先算清双 Mac Studio 的硬件预算。
- 依据：性能对比与运行配置均为 Meta 官方口径，暂无独立验证；对标对象包括 Qwen 3.8、GLM 5.3、Kimi K3 等开源模型，许可证条款出自官方说明。
After its Llama series, Meta is returning to open-weight territory with Muse Spark Open. Meta calls it the most capable US open-weight model, built for coding and agentic tasks, with 816 billion parameters (42 billion active), up to one million tokens of context, and support for text, image, video, and audio. On benchmarks like JobBench, MCP-Atlas, HLE-Diamond, and ZeroBench, Meta says it scores well ahead of Qwen 3.8, GLM 5.3, and Kimi K3 in several categories but still falls well short of Meta's paid Muse Wonder Full models.
The model is optimized for Apple's M5 Ultra and runs locally on two Mac Studios with 512 GB of unified memory each, according to Meta. Weights are on Hugging Face. The license is free to use but requires the label "Built with Muse Spark Open." Providers pulling in more than $20 million a year who want to offer the model as an API service need a separate agreement with Meta. More details are in the research blog.
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
来源：The Decoder：AI News · the-decoder.com