Meta 发布 Muse Ember：主打 1M 上下文的轻量智能体模型
Meta Releases Muse Ember, a 1M-Context Model for High-Volume Agent Work
Meta 出了个专门干活的迷你模型 Muse Ember，能调工具、写代码、啃长文档，1M 上下文很能装。跑分和 Claude Haiku 5.5 持平但单价只有三分之一，靠 API 跑智能体的团队值得看看成本账。
Meta 推出了其最小的模型 Muse Ember，定位是高频智能体工作负载，主要承担工具调用、编程子任务和文档解析三类场景。模型支持 1M（100 万）上下文窗口，可以直接处理长文档和大量调用历史。
在基准表现上，Muse Ember 在 Artificial Analysis Intelligence Index 上与 Claude Haiku 5.5 打平，而每个任务的成本约为后者的三分之一。这意味着对调用量大的智能体应用，可以用更低的单价获得接近的智能水平。
这类小模型瞄准的是 API 调用、批量解析等对延迟和成本敏感的场景，与 Claude Haiku 5.5 形成直接竞争。对需要在生产环境里跑大量智能体任务的工程团队，成本结构可能会明显改变。
Meta Releases Muse Ember, a 1M-Context Model for High-Volume Agent Work
Meta Muse Ember is Meta’s smallest model, built for tool calls, coding subtasks and document parsing. It ties Claude Haiku 5.5 on the Artificial Analysis Intelligence Index at roughly a third of the cost per task. The post Meta Releases Muse Ember, a 1M-Context Model for High-Volume Agent Work appeared first on MarkTechPost .
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Muse Ember, Meta, Claude Haiku 5.5, 智能体.
Primary source verification: published directly by marktechpost with first-party authority.
