Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Real-world Visual Question Answering on MME-RealWorld-Lite (MMERW)
Loading...
49.03
Accuracy
SSL4RL-7B (Mask)
29.4156
34.5078
39.6
44.6922
Oct 18, 2025
Oct 19, 2025
Oct 20, 2025
Oct 21, 2025
Oct 22, 2025
Oct 23, 2025
Oct 24, 2025
Accuracy
Updated 15d ago
Evaluation Results
Method
Method
Links
Accuracy
SSL4RL-7B (Mask)
Category=SSL4RL-7B, SS...
2025.10
49.03
SSL4RL-7B (Hard-Contrastive)
Category=SSL4RL-7B, SS...
2025.10
48.61
Qwen2.5-VL-7B
Category=Base
2025.10
45.59
Qwen2.5-VL 7B + NoisyGRPO
Model=Qwen2.5-VL 7B, T...
2025.10
44.6
Qwen2.5-VL 3B + SFT
Model=Qwen2.5-VL 3B, T...
2025.10
44
Qwen2.5-VL 3B + NoisyGRPO
Model=Qwen2.5-VL 3B, T...
2025.10
44
Qwen2.5-VL 7B
Model=Qwen2.5-VL 7B, T...
2025.10
43.6
Qwen2.5-VL 7B + SFT
Model=Qwen2.5-VL 7B, T...
2025.10
43.5
Qwen2.5-VL 7B + GRPO
Model=Qwen2.5-VL 7B, T...
2025.10
42.3
Qwen2.5-VL 3B
Model=Qwen2.5-VL 3B, T...
2025.10
42.1
Qwen2.5-VL 3B + GRPO
Model=Qwen2.5-VL 3B, T...
2025.10
40.8
Position
Category=SSL4RL-3B
2025.10
38.19
Jigsaw
Category=SSL4RL-3B
2025.10
35.12
Rotation
Category=SSL4RL-3B
2025.10
34.18
Qwen2.5-VL-3B
Category=Base
2025.10
32.41
Contrastive
Category=SSL4RL-3B
2025.10
30.17
Feedback
Search any
task
Search any
task