AI 导读
Meta 发布 Muse Wonder 系列前沿模型,面向个人与专业智能体工作。作者引述数据显示 Muse Wonder Full 在 Artificial Analysis Intelligence Index 上仅比 Claude Opus 5.5 低 2 分,DeepSWE v1.1 达 79.0%(对比 Astra 74.1%、Opus 74.2%),并在 SWE-Marathon 上超过两者。作者呼吁开源 Wonder,甚至 Muse Spark 1.3。
推荐理由
作者汇总了 Muse Wonder 在多项基准上与 Claude Opus 5.5、Astra 的对比数据,读者可据此评估 Meta 新模型系列的竞争力。
正文
Meta is so back?!
Muse Wonder Full:
- Artificial Analysis Intelligence Index: only 2 points behind Opus 5.5
- DeepSWE v1.1: 79.0% vs. Astra (74.1%) and Opus (74.2%)
- SWE-Marathon: beats both Astra and Opus 5.5
please open source wonder. Even Muse Spark 1.3 would be sweet!
Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.在 X 查看被引用的帖子
来源:Yuchen Jin · x.com