Our new X account is live! Follow @wizwand_team for updates

Mathematical Reasoning on MathVision (test)

71.9Accuracy

GPT5 mini

Updated 3d ago

Evaluation Results

Method	Links
GPT5 mini 2026.01		71.9
Qwen3-VL 32B 2026.01		70.2
MMFineReason-8B 2026.01		67.1
Qwen3-VL 30B-A3B 2026.01		65.7
Gemini-2.5 Flash 2026.01		64.3
Qwen3-VL 8B 2026.01		62.7
MMFineReason-4B 2026.01		61.3
MiMo-VL-7B-SFT-2508 2025.12		57.2
MiMo-VL-Miloco-7B 2025.12		54
MMR1 8B 2026.01		48.4
MMFineReason-2B 2026.01		45.3
Qwen2.5-VL-32B + AT-RL (Ours) 2026.02		44.2
OMR 7B 2026.01		43.6
Qwen2.5-VL-72B Instruct 2026.02		43.1
Qwen2.5-VL-32B + VPPO 2026.02		42.3
Gemini 2.0 Flash 2026.02		41.3
Qwen2.5-VL-32B Instruct 2026.02		38.2
HoneyBee 8B 2026.01		37.4
Innovator-VL 2026.01		34.64
Claude 3.5 Sonnet 2026.02		33.5
MiMo-VL 2026.01		32.99
Innovator-VL 2026.01		31.32
OpenAI GPT-4o 2026.02		30.6
Qwen3-VL 2026.01		30.56
MiMo-VL 2026.01		29.77
InternVL3.5 2026.01		27.11
RTWI 2026.02		25.5
SC 2026.02		24.4
LLaVA-OV 2026.01		24.11
Self-Cer. 2026.02		24
MiniCPM-V 2026.01		23.72
RTWI 2026.02		23.4
Intern-S1 2026.01		22.76
CISC 2026.02		22.7
Deepconf 2026.02		22.7
Base 2026.02		21.4
CISC 2026.02		20.1
SC 2026.02		19.8
Deepconf 2026.02		19.7
Base 2026.02		18.7
Self-Cer. 2026.02		18