Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

OVO-Bench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Streaming Video UnderstandingOVO-Bench
Real-Time Visual Perception Avg.93.2
56
Online Video UnderstandingOVO-Bench
Backward Tracing Avg.92.33
48
Backward TracingOVO-Bench
EPM92.6
41
Real-Time Visual PerceptionOVO-Bench
OCR94
41
Real-time visual perception and backward tracingOVO-Bench
Real-Time Score93.2
24
Streaming Video UnderstandingOVO-Bench 1.0 (test)
OCR89.9
21
Real-time StreamingOVO-Bench
RTVP71.3
17
Streaming Video UnderstandingOVO-Bench RealStreamEval protocol
OCR93.9
17
Video Question AnsweringOVO-Bench
Overall Accuracy65.93
17
Video Question AnsweringOVO-Bench Backward Tracing
EPM59.93
17
Video Question AnsweringOVO-Bench Real-Time Visual Perception
OCR91.95
17
Video Backward ReasoningOVO-Bench Backward
EPM61.95
14
Video Perception ReasoningOVO-Bench Perception
OCR91.95
14
Online Visual-Only Question AnsweringOVO-bench
OCR95.3
13
Forward Active RespondingOVO-Bench
REC95.5
13
Video UnderstandingOVO-Bench
OCR94
13
Online Video UnderstandingOVO-Bench (test)
RTVP69.32
13
Video UnderstandingOVO-Bench 1.0 (full)
OCR85.9
12
Proactive AlertingOvO-Bench
Precision58.77
11
Backward TracingOVO-Bench Reactive QA 1.0 (test)
EPM56.57
10
Real-time Visual PerceptionOVO-Bench Reactive QA 1.0 (test)
OCR65.1
10
Proactive Video UnderstandingOVO-Bench
FAR Time4.4
9
Recurring alertOVO-Bench
Recall33.81
9
Single-alertOVO-Bench
PA37.5
9
Streaming NarrationOVO-Bench SSR (test)
F1 Score0.1454
8
Showing 25 of 28 rows