跳转到主要内容

Muse Wonder (Max)

MetaProprietary

描述

Muse Wonder 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 accepts text, images, video, audio, and PDFs and produces text with a 1,048,576-token context window. It is available through the Meta Model API as muse-wonder-1.0 and in Muse Code.

发布日期
2026-10-12
参数规模
—
上下文长度
1.0M
支持模态
audio, image, pdf, text, video

能力雷达图

52
general
60
coding
40
reasoning
57
science
50
agents
50
multimodal

排行榜排名

领域#排名分数来源
智能体能力模型榜105
37.0
LS
代码能力榜31
91.0
AA
通用能力榜8
90.0
AA
多模态榜141
32.0
LS
科学能力18
85.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

MemoryBench85.6%自报
JobBench70.5%自报
Personal Superintelligence Index54.3%自报
Program Bench50.0%自报
FocusBench44.9%自报
AutomationBench-AA39.1%自报

Code

DeepSWE 1.178.8%自报
SWE-Marathon53.1%自报
Terminal-Bench 4.0-AA52.0%自报

Finance

Vals Finance Agent v256.6%自报

General

Artificial Analysis52.0%自报
GDP.pdf30.8%自报

Healthcare

HealthBench Professional60.5%自报

Knowledge

Harvey LAB-AA10.0%自报

Multimodal

OSWorld 2.1 (partial)72.6%自报
OSWorld 2.1 (strict)34.3%自报

AA 评测指数

(Artificial Analysis)
Lcr(Artificial Analysis)
80.7
Scicode(UIUC + Argonne National Lab (2024))
60.3
Intelligence Index(Artificial Analysis)
52.4
Terminalbench V4 0
52.0
Hle(Center for AI Safety + Scale AI (2025))
50.1

LLM Stats 分类评分

(LLM Stats (zeroeval))
Long Context
90
Productivity
70
Finance
60
Healthcare
60
Code
60
Knowledge
50
Multimodal
50
General
50
Agents
50
Tool Calling
50
Vision
50
Reasoning
40

定价

输入价格$4 / 1M tokens
输出价格$20 / 1M tokens
混合价格(3:1)$8 / 1M tokens
缓存读取价格$0.2 / 1M tokens

速度

Tokens/秒122.9
首Token延迟23.09s
首回答延迟39.36s

供应商价格排行

供应商价格排行

2 个供应商

最便宜: Meta最贵: Venice AI
供应商输入输出
1Meta主要
$4
$20
2Venice AI
$5
$25

比较该模型在不同 API 供应商之间的定价。

外部链接