List your company

Mathematical problem solving on MATH-Vision (test)

68.8AccuracyHuman
4.73621.3683854.632May 15, 2025

Evaluation results

MethodLinks
Human2025.0568.8
Claude3.5-Sonnet#Params=-2025.0537.9
GPT-4o#Params=-2025.0530.4
GPT-4-turbo#Params=-2025.0530.3
Claude3-Opus#Params=-2025.0527.1
MathCoder-VL-8B#Params=8B2025.0526.1
InternVL2-76B#Params=76B2025.0523.6
InternVL2-26B#Params=26B2025.0523.1
GPT-4V#Params=-2025.0522.8
MathCoder-VL-2B#Params=2B2025.0521.7
InternVL2-8B#Params=8B2025.0520
Gemini-1.5-Pro#Params=-2025.0519.2
Qwen2-VL#Params=8B2025.0519.2
MathGLM-Vision-9B#Params=9B2025.0519.2
MAVIS-7B#Params=7B2025.0519.2
IXC-2.5-Reward#Params=7B2025.0519
G-LLaVA-7B#Params=7B2025.0516.6
Multimath-7B#Params=7B2025.0516.3
Math-LLaVA-13B#Params=13B2025.0515.7
Qwen-VL-Max#Params=-2025.0515.6
InternVL-Chat-2B-V1-5#Params=2B2025.0515.3
IXC-2-VL#Params=7B2025.0514.5
Math-PUMA-Qwen2#Params=8B2025.0514
LLaVA-1.5-13B#Params=13B2025.0511
Qwen-VL-Plus#Params=-2025.0510.8
Random Chance2025.057.2