Qwen2.5-Omni-7B
Alibaba Cloud / Qwen TeamQwenOpen WeightApache 2.0 · Commercial OK
描述
Qwen2.5-Omni is the flagship end-to-end multimodal model in the Qwen series. It processes diverse inputs including text, images, audio, and video, delivering real-time streaming responses through text generation and natural speech synthesis using a novel Thinker-Talker architecture.
發布日期
2025-03-27
參數規模
7.0B
上下文長度
—
支援模態
—
能力雷達圖
50
general
80
coding
60
reasoning
26
science估算
0
agents
90
multimodal
Science 在缺少專門科學評測時使用推理能力代理估算。
排行榜排名
| 領域 | #排名 | 分數 | 來源 |
|---|---|---|---|
| 多模态榜 | 52 | 74.0 | LS |
基準測試分數 (LLM Stats)
Audio
VocalSound
93.9%自報
GiantSteps Tempo
88.0%自報
MMAU Music
69.2%自報
MMAU Sound
67.9%自報
MMAU
65.6%自報
MMAU Speech
59.8%自報
OmniBench Music
52.8%自報
CoVoST2 en-zh
0.41 / 100自報
MusicCaps
32.8%自報
Common Voice 15
0.08 / 100自報
Biology
GPQA
30.8%自報
Code
HumanEval
78.7%自報
Communication
VoiceBench Avg
74.1%自報
MM-MT-Bench
0.06 / 100自報
Creativity
Meld
57.0%自報
Finance
MMLU-Pro
47.0%自報
General
MBPP
0.73 / 100自報
MMLU-Redux
71.0%自報
MultiPL-E
65.8%自報
MMStar
64.0%自報
MME-RealWorld
61.6%自報
MMMU
59.2%自報
MMMU-Pro
36.6%自報
LiveBench
29.6%自報
NMOS
0.05 / 100自報
Grounding
PointGrounding
66.5%自報
Healthcare
CRPErelation
76.5%自報
Image To Text
DocVQA
95.2%自報
TextVQA
84.4%自報
OCRBench_V2
57.8%自報
Language
FLEURS
0.04 / 100自報
Long Context
EgoSchema
68.6%自報
Math
GSM8k
88.7%自報
MATH
71.5%自報
MathVista
67.9%自報
MathVision
25.0%自報
Multimodal
ChartQA
85.3%自報
AI2D
83.2%自報
MMBench-V1.1
81.8%自報
VideoMME w sub.
72.4%自報
MVBench
70.3%自報
MuirBench
59.2%自報
OmniBench
56.1%自報
Spatial Reasoning
RealWorldQA
70.3%自報
Vision
ODinW
42.4%自報
AA 評測指數
暫無 AA 評測資料
LLM Stats 分類評分
Image To Text90
Code80
Spatial Reasoning70
Video70
Vision70
Long Context70
Math60
Multimodal60
Reasoning60
Finance50
General50
Healthcare50
Language50
Legal50
Biology30
Chemistry30
Physics30
Communication10
Speech To Text0
定價
暫無定價資料
速度
暫無速度資料
可用提供商
(LS 內部計價單位)暫無提供商資料