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

Long-context video understanding on LongVideoBench (test)

66.7Accuracy

GPT-4o

Updated 1mo ago

Evaluation Results

Method	Links
GPT-4o 2024.10		66.7
GPT-4o 2024.10		66.7
GPT-4o 2024.12		66.7
ARIA 2024.10		65.3
ARIA 2024.10		65.3
Gemini-1.5-Pro 2024.10		64.4
Gemini-1.5 Pro 2024.10		64.4
LLaVA-OneVision-72B 2024.10		63.2
Gemini-1.5-Flash 2024.10		62.6
Gemini-1.5 Flash 2024.10		62.4
Qwen2-VL-72B 2024.10		61.7
GPT-4V 2024.10		60.7
InternVL2-40B 2024.10		60.6
GPT-4o mini 2024.10		58.8
GPT-4o mini 2024.10		58.8
GPT-4o mini 2024.12		58.8
NVILA 2024.12		58.7
Qwen2-VL-7B 2024.10		56.8
Qwen2-VL 2024.12		56.8
MiniCPM-V-2.6 2024.10		55.7
InternVL-Chat-V1.5 2024.10		51.7
Video-XL 2024.12		51.3
Idefics2 2024.10		49.4
Pixtral-12B 2024.10		47.4
Pixtral-12B 2024.10		47.4
Llama3.2-11B 2024.10		45.7
Llama3.2-11B 2024.10		45.7
LLaVA-NeXT-Video 2024.12		43.5