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

Real-world Multimodal Reasoning on RealWorldQA

75.4Accuracy

GPT-4o

Updated 5d ago

Evaluation Results

Method	Links
GPT-4o 2024.09		75.4	-
GPT-4o 2024.09		75.4	-
PerceptionLM-8B 2025.10		75	-
EMOVA 2024.09		71	-
MM1.5-30B 2024.09		69	-
Grok-1.5V 2024.04		68.7	-
Gemini Pro 1.5 2024.09		68.7	-
Cambrian-34B 2024.09		67.8	-
Gemini Pro 1.5 2024.04		67.5	-
InternVL 1.2 2024.04		67.5	-
EMOVA 2024.09		67.5	-
VITA 2024.09		66.8	-
InternVL 1.5 2024.04		66	-
SINKTRACK 2026.04		65.49	52.81
Gemini-1.5-Pro 2024.09		64.1	-
EMOVA 2024.09		62.6	-
Baichuan-Omni 2024.09		62.6	-
MM1.5-7B 2024.09		62.5	-
GAR-8B 2025.10		61.8	-
GPT-4V 2024.04		61.4	-
GPT-4V 2024.09		61.4	-
MM1.5-3B-MoE 2024.09		60.7	-
BLIP-3 2024.09		60.5	-
Phi-3-Vision-4B 2024.09		59.4	-
MM1-30B 2024.09		59.4	-
VITA 2024.09		59	-
GAR-8B 2025.10		58.7	-
SINKTRACK 2026.04		57.91	52.54
CoT 2026.04		57.86	43
MM1.5-1B-MoE 2024.09		57.8	-
CoT 2026.04		57.69	49.92
InternVL2-2B 2024.09		57.4	-
Direct 2026.04		57.34	50.96
Dynamic-LLaVA-7B† 2024.12		57	-
MM1.5-3B 2024.09		56.9	-
GPT-4V 2024.09		56.5	-
MiniCPM-V 2.0-3B 2024.09		55.8	-
MM1-3B 2024.09		55.8	-
MM1-7B 2024.09		55.7	-
LLaVAOne Vision-0.5B 2024.09		55.6	-
IVTP 2024.12		54.6	-
DAM-3B 2025.10		54.3	-
LLaVA-1.5-7B 2024.12		53.7	-
LLaVA-FastV 2024.12		53.7	-
MM1.5-1B 2024.09		53.3	-
SINKTRACK 2026.04		52.46	47.39
Claude-3 Sonnet 2024.04		51.9	-
Direct 2026.04		51.33	48.43
MM1-1B 2024.09		51.2	-
Claude-3 Opus 2024.04		49.8	-
CoT 2026.04		49.41	44.53
SINKTRACK 2026.04		48.28	28.42
Direct 2026.04		47.49	39.93
LLaVA-1.5-7B+H2O 2024.12		42.3	-
CoT 2026.04		35.69	26.46
Direct 2026.04		31.15	23.25
PAM-3B 2025.10		1.7	-