Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multimodal Reasoning on R1-Onevision-Bench Deduction
Loading...
27.8
Accuracy
No Compression
21.144
22.872
24.6
26.328
Feb 10, 2026
Accuracy
Average Length
Ratio Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Average Length
Ratio Score
No Compression
Base Model=Qwen2.5-VL
2026.02
27.8
834.3
30
XMCC
Base Model=Qwen2.5-VL
2026.02
27.6
115
4.2
Prune-on-Logic
Base Model=Qwen2.5-VL
2026.02
27.3
588.3
21.5
StepEntropy
Base Model=Qwen2.5-VL
2026.02
26.5
522.2
19.7
Prune-on-Logic
Base Model=Qwen3-VL
2026.02
23
602.6
26.2
XMCC
Base Model=Qwen3-VL
2026.02
22.8
103.7
4.5
StepEntropy
Base Model=Qwen3-VL
2026.02
22.5
560
24.9
No Compression
Base Model=Qwen3-VL
2026.02
21.4
819
38.3
Feedback
Search any
task
Search any
task