List your company

Visual Mathematical Reasoning on FlowVerse Vision Centric RP+EI+OQ

32.6CoT ErrorQwen-VL-Plus
31.09641.24851.461.552Mar 19, 2025

Evaluation results

MethodLinks
Qwen-VL-Plus2025.0332.626
Gemini-Pro2025.0336.127.3
InfiMM-Math-7B2025.0339.630.3
InternVL2.5-8B2025.034135.9
MathLLaVA-13B2025.0341.735.9
Qwen-VL-Max2025.0342.733.2
VLM-R1-7B†2025.0346.137.8
MultiMath-7B2025.0347.240.4
SVE-Math-Qwen2.5-7B2025.0348.944.2
Qwen2.5-VL-7B2025.0350.846.3
Qwen2-VL-72B2025.0350.942.1
MathFlowBackbone=Qwen2.5-VL-7B2025.0354.551.6
Claude-3.5-Sonnet2025.0356.548
GPT-4o-mini2025.0357.450.1
GPT-4o2025.0358.652.6
GPT-4V2025.0361.846.3
InternVL2.5-78B2025.036358.8
Gemini-2.5-pro2025.0365.757.9
GPT-52025.0367.161.7
MathFlowBackbone=GPT-52025.0370.266.7