Meta releases its first frontier model, Muse Wonder
https://techcrunch.com/2026/10/12/meta-releases-its-first-frontier-model-muse-wonder/📌 Meta 首個前沿模型登場：Muse Wonder 主打長期記憶而非只拚分數
TL;DR：Meta 發布首款前沿模型 Muse Wonder，效能看齊 Fable 5.1、GPT-6 Astra，強調跨時間記憶能力。
在 OpenAI、Anthropic 輪番推出旗艦模型的這段時間，Meta 一直缺席「前沿模型」這張考卷。這週一，Meta Superintelligence Lab 終於交卷，而交出的答案，重點不是跑分有多高，而是模型「記不記得你」。
🤔 背景：從輕量 Spark 到重量級 Wonder Meta 先前推出的 Muse Spark 定位是低延遲的輕量模型；這次新發布的 Muse Wonder 與 Muse Wonder Full 則是體量更大、能力更強的版本，在複雜任務（例如 coding）上表現更好。依 Meta 在模型卡中揭露的測試結果，基準分數可與 Fable 5.1、GPT-6 Astra 相提並論。Meta 官方部落格寫道，這兩款模型「標誌著我們在 scaling ladder 上的下一步，也是邁向個人超級智慧的重大一步」。
🧩 兩個版本，不同管道 Wonder 透過 Meta Model API 提供，每百萬 input token 收費 4 美元、output token 20 美元，與 GPT-6 Sol 或 Opus 5.5 的費率大致相當。Wonder Full 則僅開放給 Muse Code 訂閱戶使用，不會透過 API 對外提供。兩款模型同時也會用來驅動 Muse 這個 agent 產品。
💡 重點不是跑分，是「active memory」 Meta 這次特別強調模型保留、整合使用者長期提供資訊的能力，稱之為「active memory」。為了驗證這項能力，Meta 引用了挑戰持續學習（continual learning）與記憶建構框架的 MemoryBench 基準，並宣稱 Muse Wonder 在這項基準上的分數高於 Fable、Opus、Astra。研究科學家 Kavya Kopparapu 在發布影片中表示：「我們投入了大量心力，確保你的模型能在數天、數週、數月甚至更久的時間裡，記住對你重要的事。」
⚠️ 時間點值得留意 這是 Scale AI 創辦人 Alexandr Wang 去年 8 月重整 Meta AI 團隊以來，Meta 首次推出前沿等級模型。在此之前，Meta 只發布過輕量的 Muse Spark 與 Muse agent，並未像 OpenAI、Anthropic 那樣頻繁推出旗艦級模型。這次訓練地點是今年初剛啟用、位於美國俄亥俄州中部的 Meta Prometheus 資料中心。
🎯 實務啟示 對於需要長期維運個人化助理或對話系統的工程團隊而言，Muse Wonder 強調的「active memory」與 MemoryBench 這類評測，提示了一個值得關注的方向：模型能力的比較不該只看單次回答的跑分，也要看它跨會話、跨時間保留與調用資訊的穩定性。若你的產品線也需要長期記憶能力，MemoryBench 的設計思路值得留意。
🔗 來源
- 標題：Meta releases its first frontier model, Muse Wonder
- 作者／機構：Russell Brandom／TechCrunch AI
- 連結：https://techcrunch.com/2026/10/12/meta-releases-its-first-frontier-model-muse-wonder/
#Meta #MuseWonder #LLM #AIAgent #FrontierModel #AIMemory #GenerativeAI #MachineLearning #AIResearch #SuperintelligenceLab
原始資料 TechCrunch AI · 收集於 2026-10-12
摘要原文
Meta’s Superintelligence Lab released a pair of new state-of-the-art models on Monday, Muse Wonder and Muse Wonder Full, representing the company’s first attempt at a frontier model. Where Muse Spark aimed to be a lightweight model that could perform at low latency, the Wonder models are larger and said to be more capable, with higher benchmark scores and better results on complex tasks like coding. According to Meta’s testing (presented in the model card ), Wonder and Wonder Full achieve comparable benchmark scores to Fable 5.1 and GPT-6 Astra. “Together, they mark the next step on our scaling ladder and a major leap toward personal superintelligence,” Meta’s blog post about the models reads. That scale is reflected in the pricing: Accessed through the Meta Model API, Wonder will cost $4 per million input tokens and $20 per million output tokens, making it roughly equivalent in cost to GPT-6 Sol or Opus 5.5. Notably, Wonder Full will only be available to Muse Code subscribers and will not be served through the API. Both models will also be used to power the Muse agent. In keeping with Meta’s interest in developing personal assistants that can perform tasks over a long period, the company placed particular emphasis on the new models’ ability to retain and synthesize information given by the user, a property Meta calls “active memory.” To achieve better recall, Meta highlighted the MemoryBench benchmark , which challenges continual learning and memory construction frameworks. According to the company’s testing, Muse Wonder now scores higher on MemoryBench than Fable, Opus, or Astra. “We put a tremendous amount of effort into making sure your model can remember what’s important to you over days, weeks, months, and beyond,” said research scientist Kavya Kopparapu in a video accompanying the launch. The new models are the first frontier model release since Scale AI founder Alexandr Wang restructured Meta’s AI efforts in August of last year . Since then, Meta has released its lightweight Muse Spark model and Muse agent but has avoided the kind of state-of-the-art model releases OpenAI and Anthropic are known for. According to the post, the models were trained at Meta’s Prometheus data center in central Ohio , which came online earlier this year.
由 claude-code/sonnet 自動生成