Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Complex Scene Reasoning on EMMA mini
Loading...
25.25
Score
SceneAlign
13.55
16.5875
19.625
22.6625
Jan 9, 2026
Score
Updated 4d ago
Evaluation Results
Method
Method
Links
Score
SceneAlign
Model=Qwen3-VL-4B
2026.01
25.25
SFT
Model=Qwen3-VL-4B
2026.01
24
SceneAlign
Model=Qwen2.5-VL-3B
2026.01
23.25
SceneAlign
Model=InternVL3-8B
2026.01
23.25
SFT
Model=Qwen2.5-VL-3B
2026.01
22.75
Base
Model=Qwen3-VL-4B
2026.01
22.75
SceneAlign
Model=Qwen2.5-VL-7B
2026.01
22.75
SFT
Model=InternVL3-8B
2026.01
22.25
SFT
Model=Qwen2.5-VL-7B
2026.01
21
Base
Model=InternVL3-8B
2026.01
21
Base
Model=Qwen2.5-VL-3B
2026.01
20
SceneAlign
Model=LLaVA-Next-8B
2026.01
19.5
Base
Model=Qwen2.5-VL-7B
2026.01
17.75
SFT
Model=LLaVA-Next-8B
2026.01
17
AoT
Model=LLaVA-Next-8B
2026.01
16.75
Base
Model=LLaVA-Next-8B
2026.01
15.75
LLaVA-Reasoner
Model=LLaVA-Next-8B
2026.01
14
Feedback
Search any
task
Search any
task