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

Complex Reasoning on cvbench

54.9Accuracy

Qwen3-VL-4B + SynRL

53.75654.05354.3554.647Mar 18, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
54.9
2026.03
54.9
2026.03
54.1
2026.03
53.8