Share your thoughts, 1 month free Claude Pro on usSee more

Multi-modal Visual Capability on MMStar

63.9Score

GPT-4o-0513

Updated 1mo ago

Evaluation Results

Method	Links
GPT-4o-0513 2024.09		63.9
Claude3.5-Sonnet 2024.09		62.2
OneVision 2024.09		61.9
Defender Iter. 3 2026.01		61.53
InternVL2-8B 2024.09		61.5
Defender Iter. 2 2026.01		61.4
Defender Iter. 1 2026.01		61.27
Liu et al. (Insert) 2026.01		61.2
POINTS-7B 2024.09		61
Ablation (Clean Data) 2026.01		60.93
Yang et al. 2026.01		60.87
Liu et al. (All) 2026.01		60.8
Liu et al. (Add) 2026.01		60.4
Base (M_def^(0)) 2026.01		60.33
IXC-2.5 2024.09		59.9
Gemini-1.5-Pro 2024.09		59.1
Ovis1.5-LLAMA3-8B 2024.09		57.3
POINTS-9B 2024.09		56.4
Idefics3-LLAMA3-8B 2024.09		55
Cambrian-34B 2024.09		54.2
Emu2-Chat 2025.03		40.7
Qwen-VL 2025.03		32.5
Chameleon-7b 2025.03		31.1
WeGen 2025.03		27.5
LLaVA 2025.03		27.1
Kosmos-2 2025.03		24.9
Idefics 2025.03		21.6
Kosmos-G 2025.03		18.4
MiniGPT4 2025.03		16.3