Agents
AITZ_EMvom Labor angegeben
81.9%inferred family alias from qwen2.5-vl-7b (score=0.3131
AndroidWorld_SRvom Labor angegeben
25.5%inferred family alias from qwen2.5-vl-7b (score=0.3131
MobileMiniWob++_SRvom Labor angegeben
91.4%inferred family alias from qwen2.5-vl-7b (score=0.3131
Multimodal
Android Control High_EMvom Labor angegeben
60.1%inferred family alias from qwen2.5-vl-7b (score=0.3131
Android Control Low_EMvom Labor angegeben
91.4%inferred family alias from qwen2.5-vl-7b (score=0.3131
CC-OCRvom Labor angegeben
77.8%inferred family alias from qwen2.5-vl-7b (score=0.3131
InfoVQAvom Labor angegeben
82.6%inferred family alias from qwen2.5-vl-7b (score=0.3131
MMBenchvom Labor angegeben
84.3%inferred family alias from qwen2.5-vl-7b (score=0.3131
MMBench-Videovom Labor angegeben
1.8%inferred family alias from qwen2.5-vl-7b (score=0.3131
MVBenchvom Labor angegeben
69.6%inferred family alias from qwen2.5-vl-7b (score=0.3131
PerceptionTestvom Labor angegeben
70.5%inferred family alias from qwen2.5-vl-7b (score=0.3131
TempCompassvom Labor angegeben
71.7%inferred family alias from qwen2.5-vl-7b (score=0.3131
VideoMME w sub.vom Labor angegeben
71.6%inferred family alias from qwen2.5-vl-7b (score=0.3131
VideoMME w/o sub.vom Labor angegeben
65.1%inferred family alias from qwen2.5-vl-7b (score=0.3131
Language
CharadesSTAvom Labor angegeben
43.6%inferred family alias from qwen2.5-vl-7b (score=0.3131
Reasoning
Hallusion Benchvom Labor angegeben
52.9%inferred family alias from qwen2.5-vl-7b (score=0.3131
Long_context
LongVideoBenchvom Labor angegeben
54.7%inferred family alias from qwen2.5-vl-7b (score=0.3131
LVBenchvom Labor angegeben
45.3%inferred family alias from qwen2.5-vl-7b (score=0.3131
MLVUvom Labor angegeben
70.2%inferred family alias from qwen2.5-vl-7b (score=0.3131
Math
MathVisionvom Labor angegeben
25.1%inferred family alias from qwen2.5-vl-7b (score=0.3131
MathVista-Minivom Labor angegeben
68.2%inferred family alias from qwen2.5-vl-7b (score=0.3131
General
MMStarvom Labor angegeben
63.9%inferred family alias from qwen2.5-vl-7b (score=0.3131
MMT-Benchvom Labor angegeben
63.6%inferred family alias from qwen2.5-vl-7b (score=0.3131
MMVetvom Labor angegeben
67.1%inferred family alias from qwen2.5-vl-7b (score=0.3131
Image_to_text
OCRBenchvom Labor angegeben
86.4%inferred family alias from qwen2.5-vl-7b (score=0.3131
TextVQAvom Labor angegeben
84.9%inferred family alias from qwen2.5-vl-7b (score=0.3131
Sonstige
ChartQAvom Labor angegeben
87.3%inferred family alias from qwen2.5-vl-7b (score=0.3131
DocVQAvom Labor angegeben
95.7%inferred family alias from qwen2.5-vl-7b (score=0.3131
58.6%inferred family alias from qwen2.5-vl-7b (score=0.3131
MMMU Provom Labor angegeben
38.3%inferred family alias from qwen2.5-vl-7b (score=0.3131
ScreenSpotvom Labor angegeben
84.7%inferred family alias from qwen2.5-vl-7b (score=0.3131
ScreenSpot-Provom Labor angegeben
29%inferred family alias from qwen2.5-vl-7b (score=0.3131
Laborangaben stammen aus den Veröffentlichungsunterlagen des jeweiligen Labors und können andere Einstellungen nutzen, daher wird keine davon als beste markiert. Unabhängige Ergebnisse stammen von externen Prüfern.
