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

OmniBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Vision-Audio-TextOmnibench
Accuracy47.6
34
Audio-Visual ReasoningOmniBench
Accuracy63.6
16
Humanoid Motion TrackingOmniBench Walk (Slow)
Success Rate (SR)100
14
Loco-ManipOmniBench Low 1.0 (test)
Success Rate (SR)100
14
Loco-ManipOmniBench Medium 1.0 (test)
Success Rate (SR)100
14
Loco-ManipOmniBench High 1.0 (test)
Success Rate (SR)100
14
Omni-modal reasoningOmniBench (test)
Score43.6
12
Safety EvaluationOmniBench
Accuracy42.47
12
Integrated Multimodal ReasoningOmniBench (held-out)
Causal Score55.2
11
Audio-video understandingOmniBench
Score49.1
10
Multimodal ReasoningOmniBench
Causal Score54.9
9
Omni-modal UnderstandingOmniBench
Overall Score58.41
8
Humanoid Motion TrackingOmniBench Jump (Low)
Success Rate100
7
Humanoid Motion TrackingOmniBench Jump Medium
Success Rate100
7
Humanoid Motion TrackingOmniBench Jump (High)
Success Rate (SR)90
7
Humanoid Motion TrackingOmniBench Run Medium
Success Rate (SR)100
7
Humanoid Motion TrackingOmniBench Run (Fast)
Success Rate (SR)100
7
Humanoid Motion TrackingOmniBench Walk (Fast)
Success Rate100
7
SquatOmniBench Low 1.0 (test)
Success Rate (SR)100
7
SquatOmniBench Medium 1.0 (test)
SR100
7
SquatOmniBench High 1.0 (test)
Success Rate (SR)100
7
Audio-Visual QAOmniBench
Accuracy48.25
6
Image Question AnsweringOmniBench
Metric-
0
Showing 23 of 23 rows