List your company

Visual Mathematical Reasoning on MathVista 1.0 (testmini)

67.6AccuracyQwen2.5-VL-7B
44.7250.6656.662.54Mar 3, 2026

Evaluation results

MethodLinks
Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2026.0367.6
Qwen2.5-VL-7B + SFT+GRPOBackbone=Qwen2.5-VL-7B, Stage=SFT+GRPO2026.0366.5
Qwen2.5-VL-7B + SFTBackbone=Qwen2.5-VL-7B, Stage=SFT2026.0366.3
Qwen2.5-VL-7B + SFT+DPOBackbone=Qwen2.5-VL-7B, Stage=SFT+DPO2026.0365.5
Qwen2.5-VL-7B + SFT+DPO+GRPOBackbone=Qwen2.5-VL-7B, Stage=SFT+DPO+GRPO2026.0365.2
Qwen2.5-VL-7B + SPA-VLBackbone=Qwen2.5-VL-7B, Alignment=SPA-VL2026.0364
Qwen2.5-VL-7B + SaFeR-VLMBackbone=Qwen2.5-VL-7B, Alignment=SaFeR-VLM2026.0363.9
Qwen2.5-VL-3B + SFTBackbone=Qwen2.5-VL-3B, Stage=SFT2026.0355.3
Qwen2.5-VL-3BBackbone=Qwen2.5-VL-3B2026.0354.3
Qwen2.5-VL-3B + SFT+DPOBackbone=Qwen2.5-VL-3B, Stage=SFT+DPO2026.0354
Qwen2.5-VL-3B + VLGuardBackbone=Qwen2.5-VL-3B, Alignment=VLGuard2026.0353.8
Qwen2.5-VL-3B + SPA-VLBackbone=Qwen2.5-VL-3B, Alignment=SPA-VL2026.0353.3
Qwen2.5-VL-7B + TISBackbone=Qwen2.5-VL-7B, Alignment=TIS2026.0353.2
Qwen2.5-VL-7B + VLGuardBackbone=Qwen2.5-VL-7B, Alignment=VLGuard2026.0352.7
Qwen2.5-VL-3B + SaFeR-VLMBackbone=Qwen2.5-VL-3B, Alignment=SaFeR-VLM2026.0352.6
Qwen2.5-VL-3B + SFT+GRPOBackbone=Qwen2.5-VL-3B, Stage=SFT+GRPO2026.0352
Qwen2.5-VL-3B + SFT+DPO+GRPOBackbone=Qwen2.5-VL-3B, Stage=SFT+DPO+GRPO2026.0351.9
Qwen2.5-VL-3B + TISBackbone=Qwen2.5-VL-3B, Alignment=TIS2026.0345.6