List your company
ResearchFormerly WizWand

Scientific Reasoning on Lens

56.2AccuracyGemini2.5-Pro
28.32835.56442.850.036May 21, 2025
Updated 6mo ago

Evaluation results

MethodLinks
Gemini2.5-ProModel size=-2025.0556.2
Qwen2.5-VLModel size=72B2025.0553.65
Qwen2.5-VLModel size=32B2025.0551.66
GPT-4oModel size=-2025.0551.14
QVQ-MaxModel size=72B2025.0550.8
InternVL3Model size=78B2025.0549.39
InternVL3Model size=38B2025.0547.18
Qwen2.5-VLModel size=7B2025.0546.28
InternVL3Model size=9B2025.0544.69
Deepseek-VL2Model size=MoE 4.5B/27B2025.0544.58
Gemma3Model size=12B2025.0543.33
InternVL3Model size=2B2025.0540.56
Qwen2.5-VLModel size=3B2025.0540.33
Gemma3Model size=4B2025.0539.53
Deepseek-VL2-tinyModel size=MoE 1B/3B2025.0538.97
Kimi-VL-thinkingModel size=MoE 2.8B/16B2025.0529.4