Qwen2.5 VL 72B Instruct
Alibaba Cloud / Qwen TeamQwenOpen Weighttongyi-qianwen
Description
Qwen2.5-VL is the new flagship vision-language model of Qwen, significantly improved from Qwen2-VL. It excels at recognizing objects, analyzing text/charts/layouts in images, acting as a visual agent, understanding long videos (over 1 hour) with event pinpointing, performing visual localization (bounding boxes/points), and generating structured outputs from documents.
Release Date
2025-01-26
Parameters
72.0B
Context Length
32K
Modalities
image, text
Capability Radar
50
general
0
coding
60
reasoning
60
scienceest.
40
agents
80
multimodal
Science uses a reasoning proxy when dedicated science benchmarks are unavailable.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Agents & Tools | 82 | 45.0 | LS |
| Multimodal Ranking | 54 | 73.0 | LS |
| Reasoning | 74 | 55.0 | LS |
Benchmark Scores (LLM Stats)
Agents
AITZ_EM
83.2%SR
MobileMiniWob++_SR
68.0%SR
AndroidWorld_SR
35.0%SR
OSWorld
8.8%SR
General
MMVet
76.2%SR
MLVU-M
74.6%SR
MMStar
70.8%SR
MMMU
70.2%SR
MMMU-Pro
51.1%SR
Grounding
ScreenSpot
87.1%SR
ScreenSpot Pro
43.6%SR
Image To Text
DocVQA
96.4%SR
OCRBench
88.5%SR
OCRBench-V2 (en)
61.5%SR
Long Context
EgoSchema
76.2%SR
LVBench
47.3%SR
Math
MathVista-Mini
74.8%SR
MathVision
38.1%SR
Multimodal
Android Control Low_EM
93.7%SR
ChartQA
89.5%SR
AI2D
88.4%SR
MMBench
88.0%SR
CC-OCR
79.8%SR
TempCompass
74.8%SR
VideoMME w/o sub.
73.3%SR
PerceptionTest
73.2%SR
MVBench
70.4%SR
Android Control High_EM
67.4%SR
MMBench-Video
2.0%SR
Reasoning
Hallusion Bench
55.2%SR
AA Evaluation Indices
No AA evaluation data available
LLM Stats Category Scores
Structured Output80
Text-to-image80
Image To Text80
Spatial Reasoning70
Grounding70
Healthcare70
Reasoning70
Vision60
Long Context60
Math60
Multimodal60
Video50
General50
Agents40
Pricing
Input Price$0.25 / 1M tokens
Output Price$0.75 / 1M tokens
Blended Price (3:1)$0.375 / 1M tokens
Speed
No speed data available
Available Providers
(LS internal units)No provider data available