跳到正文
K = 0.731订阅
模型

Alexandr Wang Says Muse Ember Matches Haiku at Lower Cost and Outperforms Luna

Alexandr Wang · X· @alexandr_wang·· 3 天前

1.5万 次浏览 · 218 赞 · 10 转发 （收录时）

在 X 查看原推
原文

中文版本尚未就绪 · 显示原文

发生了什么

Alexandr Wang says Muse Ember is now on the efficient frontier 🤗 He says it matches Haiku at 3x cheaper cost per task, and is smarter than Luna.

接下来读什么

查看这件事的完整进展Meta 发布 Muse Ember，并称 Crouton Alpha 是早期版本

2 篇报道

    正文 · 中文版本尚未就绪（原文）

    our teeny model muse ember is now on the efficient frontier 🤗

    matching haiku at 3x cheaper cost per task, and smarter than luna

    引用Artificial Analysis@ArtificialAnlys

    Meta's newest model Muse Ember scores 36 on the Artificial Analysis Intelligence Index, matching Claude Haiku 5.5 (max) at a third of the Cost per Task

    Muse Ember is Meta's newest efficient-tier model line, with the same input and output token prices as Claude Haiku 5.5 and GPT-6 Luna

    Key Takeaways:

    ➤ On the Intelligence Index vs. Cost per Task Pareto frontier at both max and xhigh effort. Muse Ember (max) is the cheapest model scoring 36 or higher, at $0.14 per task, and Muse Ember (xhigh) scores 34 for $0.12. Claude Haiku 5.5 (max) costs almost 3x more at $0.41 for the same score as Muse Ember (max). Muse Ember (max) is also about 4x cheaper than DeepSeek V4.1 Flash 1017 (max, 33, $0.60) and GLM-5.3-Flash (28, $0.58)

    ➤ 8 points behind Muse Wonder for about 1/50th of the Cost per Task. Muse Wonder (max) scores 44 at $6.95 per Intelligence Index task

    ➤ Stands out on AutomationBench-AA. Muse Ember (max) scores 68.1% on AutomationBench-AA, level with Muse Wonder (max), and beats Claude Haiku 5.5 (max) on GDP.pdf (22.8% vs 20.8%), but trails it on AA-Briefcase (1521 vs 1574 Elo)

    ➤ Mixed in agentic coding in our Intelligence Index evaluations. It beats Claude Haiku 5.5 (max) on Terminal-Bench 4.0 (40.9% vs 32.8%) but trails it on FrontierCode 2.0 (27.2% vs 34.9%)

    ➤ Ahead of similarly priced models when running in its own coding agent. In Muse Code, Muse Ember (max) scores 46 on the Artificial Analysis Coding Agent Index, ahead of Claude Haiku 5.5 in Claude Code (max, 37) and GPT-6 Luna in Codex (max, 31)

    ➤ Knowledge is the main weakness, but it rarely hallucinates. Its AA-Omniscience accuracy of 21.1% is the lowest of the similarly intelligent models we compared, against 34.7% for Claude Haiku 5.5 (max) and 50.5% for Gemini 3.8 Flash (high). This is mostly because it declines to answer 64% of questions, which keeps its hallucination rate at 17.4%, half of Claude Haiku 5.5’s (max)

    Other details:
    ➤ Pricing: $0.10 input / $0.004 cached input / $0.50 output per 1M tokens
    ➤ Reasoning effort: results shown at max effort, with xhigh also on the Intelligence vs. Cost chart ➤ Availability: Meta's first-party API and Muse Code

    在 X 查看被引用的帖子

    这件事到目前为止

    Meta 发布了用于日常智能体任务的 Muse Ember。Alexandr Wang 表示，Crouton Alpha 是该模型的早期版本。Wang 称，Meta 在发布前修复了问题，并在测试中处理了超过 5 万亿个 token。 Ember 可处理工具调用、小型编码任务、路由、信息提取和长文档处理等任务。它支持文本、图像、音频和视频，上下文窗口最长可达 100 万个 token。开发者可通过 Meta Model API 和 Muse Code 使用该模型，输入价格为每百万个 token 0.10 美元，输出价格为每百万个 token 0.50 美元。 Artificial Analysis 在其 Intelligence Index 上给 Ember (max) 打出 36 分，平均每项任务成本为 0.14 美元；相比之下，Claude Haiku 5.5 (max) 每项任务成本为 0.41 美元。Ember 的编码测试结果不一：它在 Terminal-Bench 4.0 上胜过 Haiku 5.5，但在 FrontierCode 2.0 上得分较低。

    这件事的进展 · 3

    查看完整事件
    1. 11月10日 04:36
      Meta 推出 Muse Ember，面向常规智能体任务的小型模型The Decoder
    2. 11月10日 03:47
      Meta 发布 Muse Ember：其最小模型，并附基准测试成本对比宝玉 · X
    3. 11月10日 02:47

      Alexandr Wang Says Muse Ember Matches Haiku at Lower Cost and Outperforms Luna 本条

    Meta / Llama推理能力评测基准