Muse Spark
MetaProprietary
描述
Muse Spark is the first model in the Muse family developed by Meta Superintelligence Labs. It is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration. It features a Contemplating mode that orchestrates multiple agents reasoning in parallel. It demonstrates competitive performance in multimodal perception, reasoning, health, and agentic tasks, with Contemplating mode achieving 58% on Humanity's Last Exam and 38% on FrontierScience Research.
發布日期
2026-04-08
參數規模
—
上下文長度
—
支援模態
—
能力雷達圖
33
general
59
coding
88
reasoning
74
science
80
agents
70
multimodal
排行榜排名
基準測試分數 (LLM Stats)
(LLM Stats (zeroeval))Communication
Tau2 Telecom
91.5%自報
Healthcare
MedXpertQA
78.4%自報
HealthBench Hard
42.8%自報
Reasoning
GPQANYU + Cohere + Anthropic (2023)
89.5%自報
CharXiv-R
86.4%自報
IPhO 2025
82.6%自報
LiveCodeBench Pro
0.80 / 3000自報
SWE-Bench Verified
77.4%自報
Terminal-Bench 2.0Stanford × Laude Institute (2026)
59.0%自報
Humanity's Last Exam
58.4%自報
SWE-Bench ProPrinceton NLP (2024)
52.4%自報
ARC-AGI v2
42.5%自報
FrontierScience Research
38.3%自報
Search
DeepSearchQA
74.8%自報
Vision
ScreenSpot Pro
84.1%自報
MMMU-Pro
80.4%自報
SimpleVQA
0.71 / 100自報
ERQA
64.7%自報
ZEROBench
0.33 / 100自報
AA 評測指數
(Artificial Analysis)Tau2(Sierra + U Toronto + Vector Institute (2025))91.5
Gpqa(NYU + Cohere + Anthropic (2023))88.4
Lcr(Artificial Analysis)78.0
Ifbench(Google Research (2023))75.9
Terminalbench V2 162.2
Coding Index(Artificial Analysis)58.6
Terminalbench Hard(Stanford × Laude Institute (2026))45.5
Hle(Center for AI Safety + Scale AI (2025))40.7
Intelligence Index(Artificial Analysis)31.3
LLM Stats 分類評分
(LLM Stats (zeroeval))Physics90
Biology90
Chemistry90
Communication90
Frontend Development80
Grounding80
Tool Calling80
Image To Text70
Multimodal70
Reasoning70
Search70
General70
Code70
Vision70
Math60
Spatial Reasoning60
Healthcare60
Agents60
Science40
定價
輸入價格免費
輸出價格免費
混合價格(3:1)免費
速度
Tokens/秒0.0
首Token延遲0.00s
首回答延遲0.00s
供應商價格排行
暫無提供商資料