Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Complex Reasoning on Video-TT
Loading...
46.5
Accuracy
OmniJigsaw (CMM)
36.412
39.031
41.65
44.269
Mar 18, 2026
Mar 21, 2026
Mar 25, 2026
Mar 29, 2026
Apr 1, 2026
Apr 5, 2026
Apr 9, 2026
Accuracy
Updated 8d ago
Evaluation Results
Method
Method
Links
Accuracy
OmniJigsaw (CMM)
Inference Mode=w/o Audio
2026.04
46.5
OmniJigsaw (CMM)
Inference Mode=w/ Audio
2026.04
46.1
OmniJigsaw (SMS)
Inference Mode=w/o Audio
2026.04
45.9
OmniJigsaw (SMS)
Inference Mode=w/ Audio
2026.04
45.8
VideoJigsaw
Inference Mode=w/ Audio
2026.04
44.9
VideoJigsaw
Inference Mode=w/o Audio
2026.04
44.9
OmniJigsaw (JMI)
Inference Mode=w/o Audio
2026.04
44.8
OmniJigsaw (JMI)
Inference Mode=w/ Audio
2026.04
44.7
Qwen3-Omni-30B
Inference Mode=w/ Audio
2026.04
44.3
Qwen3-Omni-30B
Inference Mode=w/o Audio
2026.04
43.8
Qwen3-VL-8B + SynRL
Parameters=8B, Trainin...
2026.03
41.5
Qwen3-VL-4B + SynRL
Parameters=4B, Trainin...
2026.03
40.7
Qwen3-VL-8B
Parameters=8B
2026.03
40.6
Video-R1
Inference Mode=w/o Audio
2026.04
40.6
HumanOmniV2
Inference Mode=w/ Audio
2026.04
40.3
Qwen3-VL-4B
Parameters=4B
2026.03
38.9
HumanOmniV2
Inference Mode=w/o Audio
2026.04
38.2
Omni-R1
Inference Mode=w/o Audio
2026.04
37
Omni-R1
Inference Mode=w/ Audio
2026.04
36.8
Feedback
Search any
task
Search any
task