List your company

Multimodal Math Reasoning on V-Math

53AccuracyAutoTool (Qwen3-8B)
18.1627.20536.2545.295Dec 15, 2025

Evaluation results

MethodLinks
AutoTool (Qwen3-8B)Size=8B, Backbone=Qwen32025.1253
AutoTool (Qwen2.5-VL-7B)Size=7B, Backbone=Qwen2.5-VL2025.1244.3
GPT4oSize=-2025.1241.4
Qwen2.5-VL-72B-InstructSize=72B2025.1224.5
v-ToolRLSize=2B2025.1219.5