Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Real-World Understanding on V* Bench
Loading...
85.6
Accuracy
DeepEyes
57.52
64.81
72.1
79.39
Nov 7, 2025
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
DeepEyes
Tool=Crop, Param Size=7B
2025.11
85.6
DeepEyes
Tool=Crop, Param Size=7B
2025.11
85.6
Pixel-Reasoner
Tool=Crop, Param Size=7B
2025.11
84.3
Pixel-Reasoner
Tool=Crop, Param Size=7B
2025.11
84.3
Thyme
Tool=Code, Param Size=7B
2025.11
82.2
Thyme
Tool=Code, Param Size=7B
2025.11
82.2
DeepEyesV2
Tool=General, Param Si...
2025.11
81.8
DeepEyesV2
Tool=General, Param Si...
2025.11
81.8
InternVL3
Tool=✗, Param Size=8B
2025.11
81.2
InternVL3
Tool=✗, Param Size=8B
2025.11
81.2
Qwen2.5-VL
Tool=✗, Param Size=32B
2025.11
80.6
Qwen2.5-VL
Tool=✗, Param Size=32B
2025.11
80.6
Gemini 2.5 Pro
Tool=Code, Param Size=-
2025.11
79.6
Qwen2.5-VL
Tool=✗, Param Size=7B
2025.11
78.5
Qwen2.5-VL
Tool=✗, Param Size=7B
2025.11
78.5
LLaVA-OV
Tool=✗, Param Size=7B
2025.11
75.4
LLaVA-OV
Tool=✗, Param Size=7B
2025.11
75.4
GPT-4o
Tool=Code, Param Size=-
2025.11
58.6
Feedback
Search any
task
Search any
task