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

STEM Reasoning on LogicVista

70.02Accuracy

CodePercept-32B-S1

Updated 1mo ago

Evaluation Results

Method	Links
CodePercept-32B-S1 2026.03		70.02
Gemini2.5-Pro 2026.03		66.44
Qwen3-VL-32B-Instruct 2026.03		66.44
CodePercept-32B-S1 2026.03		65.33
CodePercept-4B-S1 2026.03		64.65
CodePercept-8B-S1 2026.03		62.64
Qwen3-VL-32B-Instruct 2026.03		62.19
CodePercept-8B-S1 2026.03		61.52
CodePercept-4B-S1 2026.03		60.4
Qwen3-VL-235A22B-Instruct 2026.03		59.73
Claude-Opus 4.1-Thinking 2026.03		59.28
Qwen3-VL-8B-Instruct 2026.03		58.16
Qwen3-VL-8B-Instruct 2026.03		56.82
Qwen3-VL-4B-Instruct 2026.03		56.82
Qwen3-VL-4B-Instruct 2026.03		54.14
Qwen3-VL-30A3B-Instruct 2026.03		53.69
GPT5-Thinking 2026.03		53.02
MiniCPM-V-4.5 2026.03		52.57
Qwen2.5-VL-72B 2026.03		52.34
InternVL3.5-8B 2026.03		51.68
Intern-S1-8B 2026.03		51.23
GLM-4.1V-9B 2026.03		51
KeyeVL1.5-8B 2026.03		45.19