メインコンテンツへスキップ

Muse Wonder Full (Max)

MetaProprietary

説明

Muse Wonder Full is a proprietary reasoning model from Meta Superintelligence Labs, released on October 12, 2026 for personal and professional agents and long-running coding tasks. Active Memory retains relevant context across interactions, and Active Focus supports sustained task execution and error recovery. Muse Wonder Full is the higher-capability variant, available to Muse Code Power Usage and Maximum Usage subscribers. It is not available through the Meta Model API. The model card specifies text, image, video, and audio inputs, text output, and a 1M-token inference context for both variants.

リリース日
2026-10-12
パラメータ
—
コンテキスト長
—
モダリティ
—

能力レーダー

55
general
63
coding
50
reasoning
59
science
50
agents
50
multimodal

ランキング

ベンチマークスコア (LLM Stats)

(LLM Stats (zeroeval))

Agents

MemoryBench85.5%自己申告
JobBench75.3%自己申告
Personal Superintelligence Index59.1%自己申告
Program Bench59.0%自己申告
FocusBench54.2%自己申告
AutomationBench-AA45.4%自己申告

Code

DeepSWE 1.179.0%自己申告
SWE-Marathon57.5%自己申告
Terminal-Bench 4.0-AA57.1%自己申告

Finance

Vals Finance Agent v259.1%自己申告

General

Artificial Analysis56.0%自己申告
GDP.pdf33.2%自己申告

Healthcare

HealthBench Professional62.3%自己申告

Knowledge

Harvey LAB-AA13.1%自己申告

Multimodal

OSWorld 2.1 (partial)78.4%自己申告
OSWorld 2.1 (strict)39.4%自己申告

AA評価指数

(Artificial Analysis)
Lcr(Artificial Analysis)
83.7
Scicode(UIUC + Argonne National Lab (2024))
62.6
Terminalbench V4 0
57.1
Intelligence Index(Artificial Analysis)
55.8
Hle(Center for AI Safety + Scale AI (2025))
53.6

LLM Statsカテゴリスコア

(LLM Stats (zeroeval))
Long Context
90
Productivity
80
Finance
60
General
60
Healthcare
60
Agents
60
Code
60
Knowledge
50
Multimodal
50
Reasoning
50
Tool Calling
50
Vision
50

価格設定

入力価格無料
出力価格無料
混合価格(3:1)無料

速度

トークン/秒0.0
初トークン遅延0.00s
初回答遅延0.00s

プロバイダー価格ランキング

プロバイダーデータがありません

外部リンク