Meta Launches Muse Wonder and Muse Wonder Full: Benchmarks, Pricing and How to Use It
https://www.marktechpost.com/2026/10/13/meta-muse-wonder-muse-wonder-full-release/📌 Meta 發布 Agent 旗艦模型 Muse Wonder，定價與操作流程一次看懂
TL;DR：Meta 推出新一代 agent 模型 Muse Wonder／Muse Wonder Full，可規劃長任務並操作瀏覽器、桌面與手機 App，API 已開放申請。
多數人對「AI agent」的想像還停在「幫你寫一段程式」，但 Meta 這次推出的 Muse Wonder，直接把目標設定在「長時間自主作業」：寫程式、處理法務與財務任務，甚至操作電腦。
🤔 接續 Muse Spark 的 scaling ladder
Muse Wonder 是 Meta Muse 系列在 scaling ladder 上接續 Muse Spark 的下一步。Meta 團隊表示模型在自家的 Prometheus 資料中心訓練，pretraining 與 reinforcement learning 兩端都有提升，並將表現進步歸功於更大的 base model，使模型在陌生任務上的泛化能力更好。
🧩 核心能力：長任務規劃＋操作電腦
對開發者而言，Muse Wonder 的重點能力是規劃長任務並把子任務委派給 sub-agent，同時具備操作瀏覽器、桌面應用程式與手機 App 的能力，目標是讓 agent 能完成需要多步驟、跨工具的實際工作，而不只是單輪問答。Muse Wonder 已在 Meta Model API 與 Muse Code 上線；能力更完整的 Muse Wonder Full 則限定付費 Muse Code 方案才能使用。
📊 Benchmark 數字揭露有限，比較方式也不完全公平
Meta 官方表格中列出的分數多半引用第三方來源，例如 Artificial Analysis、Vals AI 與 Datacurve，而 PSI、MemoryBench、FocusBench 則是 Meta 內部評測項目。值得注意的是，model card 本身就說明競品的跑分是「best-effort」，不保證對手模型有被妥善調校；SWE-Marathon 項目中，對手使用的是 Codex 與 Claude Code 跑出的結果，並非同一套測試環境下的比較。由於素材並未提供實際分數，這裡不列出具體數字，但這個揭露限制本身就值得工程師在解讀官方 benchmark 時多留一分警覺。
🧩 怎麼開始用
- 打開 dev.meta.ai，點選 Start building 登入；Muse Code 的安裝指令也在同一頁。
- 選模型：Muse Wonder 定價為每百萬 token 輸入 $4、cached $0.20、輸出 $20；Muse Wonder Full 需要 Muse Code 的 Power（$50/月）或 Maximum（$200/月）方案。
- 用 Meta 帳號登入，確認名稱並接受 Model API 條款。
- 加入付款方式後，在 API keys 頁面建立 API key，並依 quickstart 指示匯出為
MODEL_API_KEY。 - 在 Playground 的 Chat 頁面選擇
muse-wonder-1.0，可調整 temperature、max tokens 與 reasoning effort。 - 點 Get code 取得 curl 或 Python 範例，呼叫的是 Responses endpoint
https://api.meta.ai/v1/responses；此 API 與 OpenAI 相容，換個 base URL 就能直接用 OpenAI SDK。 - 也可以在終端機用 Muse Code：Mac 上執行
curl -fsSL https://dev.meta.ai/install.sh | bash，Windows 用irm https://dev.meta.ai/install.ps1 | iex，在專案資料夾執行muse並登入。預設模型是muse-spark-1.2，可用/models切換，或直接muse --model muse-wonder-1.0啟動。
⚠️ 解讀 benchmark 時要留意比較基準不同
由於競品跑分未必經過調校、且部分比較（如 SWE-Marathon）用的是不同 harness，建議工程師自行用實際任務測試，而非只看官方發布的排行榜。
🎯 實務啟示
對已經在用 OpenAI SDK 的團隊，Muse Wonder 的相容 API 讓遷移成本很低，可以先用 Muse Wonder（較低成本）驗證 agent 工作流，再視需求評估是否要升級到需要付費方案的 Muse Wonder Full。
🔗 來源
- 標題：Meta Launches Muse Wonder and Muse Wonder Full: Benchmarks, Pricing and How to Use It
- 作者／機構：Michal Sutter, MarkTechPost
- 連結：https://www.marktechpost.com/2026/10/13/meta-muse-wonder-muse-wonder-full-release/
#Meta #MuseWonder #AIAgent #LLM #AgenticAI #ComputerUse #MetaAI #DeveloperTools #APIIntegration #AIModels
原始資料 MarkTechPost · 收集於 2026-10-14
摘要原文
Meta has released Muse Wonder and Muse Wonder Full , its most capable models so far. Muse Wonder is a frontier model from Meta built for personal and professional AI agents. Both models target long-running agent work: coding, legal and finance tasks, and computer use. Muse Wonder is live now on the Meta Model API and Muse Code . Muse Wonder Full is limited to paid Muse Code plans. Muse Wonder follows Muse Spark on Meta’s scaling ladder. Meta team states that it was trained at its Prometheus data center , with gains across pretraining and reinforcement learning. Meta credits a larger base model for better results on unfamiliar tasks. Two named behaviors define the release: For developers, Muse Wonder plans long tasks and delegates to sub-agents. It also operates browsers, desktop and mobile apps. All scores below come from Meta’s published table at max reasoning effort. Several rows use third-party sources such as Artificial Analysis, Vals AI and Datacurve. Note: PSI, MemoryBench and FocusBench are internal Meta evaluations. The model card says competitor runs are best-effort and may not be tuned. SWE-Marathon results for rivals used Codex and Claude Code, so it is not a common-harness comparison. Step 1: Open dev.meta.ai. Click Start building to sign in. The Muse Code install command sits on the same page. Step 2: Pick your model. Muse Wonder costs $4 input, $0.20 cached and $20 output per million tokens. Muse Wonder Full needs a Muse Code plan : Power at $50 or Maximum at $200 per month. Step 3: Sign in with your Meta account. Confirm your name, accept the Model API terms, and click Get started . Step 4. Add billing, then create an API key. New accounts see a “No payment method” banner, and Create API key stays disabled. Add a card under Billing , then open API keys and select Create API key . Export it as MODEL_API_KEY , per the quickstart . Step 5. Test it in the Playground. Open Chat , pick muse-wonder-1.0 , and set temperature, max tokens and reasoning effort. Step 6. Copy working code. Click Get code for a curl or Python snippet. It calls the Responses endpoint at https://api.meta.ai/v1/responses . The API is OpenAI-compatible, so the OpenAI SDK works with a new base URL: Step 7: Or use Muse Code in your terminal. On Mac, install with curl -fsSL https://dev.meta.ai/install.sh | bash . On Windows, use irm https://dev.meta.ai/install.ps1 | iex . Run muse in your project folder, trust the workspace, and sign in. The default model is muse-spark-1.2 . Switch with /models , or launch with muse --model muse-wonder-1.0 . Check out the research blog , model card and developer docs . All credit goes to the researcher of this project. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter . Wait! are you on telegram? now you can join us on telegram as well. The post Meta Launches Muse Wonder and Muse Wonder Full: Benchmarks, Pricing and How to Use It appeared first on MarkTechPost .
由 claude-code/sonnet 自動生成