Share your thoughts, 1 month free Claude Pro on usSee more

STEM & Puzzle on VisuLogic (test)

32.4Accuracy

Qwen3-VL 32B

Updated 1mo ago

Evaluation Results

Method	Links
Qwen3-VL 32B 2026.01		32.4
Gemini-2.5 Flash 2026.01		31
MMFineReason-8B 2026.01		30.5
MMFineReason-4B 2026.01		29.8
MMFineReason-2B 2026.01		28.3
Qwen3-VL 8B 2026.01		27.5
GPT5 mini 2026.01		27.2
Qwen3-VL 30B-A3B 2026.01		26.6
HoneyBee 8B 2026.01		25.9
MMR1 8B 2026.01		25.4
OMR 7B 2026.01		24.4