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

Video Understanding on EgoSchema (test)

77.9Accuracy

Qwen2-VL-72B

Updated 29d ago

Evaluation Results

Method	Links
Qwen2-VL-72B 2024.09		77.9	-
GPT-5 2026.01		75.6	-
Claude Sonnet 4.5 2026.01		73.1	-
GPT-4o 2024.09		72.2	-
Gemini 2.5 Pro 2026.01		72.2	-
Eagle2.5-8B 2026.01		72.2	-
Gemini-1.5-Pro 2025.03		72.2	-
GPT-5 mini 2026.01		70.9	-
Gemini 2.5 Flash 2026.01		70.2	-
Qwen3-VL-8B 2026.01		69.8	-
Gemini 3 Pro 2026.01		68.9	-
PLM-8B 2026.01		68.8	-
Qwen3-VL-4B 2026.01		68.4	-
PLM-3B 2026.01		66.9	-
Qwen2-VL-7B 2024.09		66.7	-
Qwen2-VL-7B-Ins. 2024.12		66.7	-
Gemini-1.5-Flash 2025.03		65.7	-
Vanilla 2024.12		65	100
iLLaVA 2024.12		63.3	97.4
Gemini 1.5-Pro 2024.09		63.2	-
iLLaVA 2024.12		62.7	96.4
GLM-4.1V-9B 2026.01		62.6	-
VisionZip 2024.12		62.5	96.2
FasterVLM 2024.12		62.3	95.9
PyramidDrop 2024.12		62.2	95.8
Previous SoTA 2024.09		62	-
LLaVA-OV-72B 2024.12		62	-
Molmo2-8B 2026.01		62	-
FasterVLM 2024.12		61.8	95
VisionZip 2024.12		61.7	94.9
PyramidDrop 2024.12		61.5	94.6
Molmo2-4B 2026.01		61.2	-
SparseVLM 2024.12		61.1	94
SparseVLM 2024.12		60.5	93
LLaVA-OV-7B 2024.12		60.1	-
InternVL3.5-4B 2026.01		58.9	-
LLaVA-OV-72B 2024.12		58.6	-
InternVL3.5-8B 2026.01		58.6	-
MAmmoTH-VL-8B 2024.12		58.5	-
LLaVA-Video-7B 2026.01		57.3	-
Molmo2-O-7B 2026.01		56.8	-
Keye-VL-1.5-8B 2026.01		56.3	-
Qwen2-VL-2B 2024.09		54.9	-
VideoChat2-7B 2025.03		54.4	-
InternVL-2-8B 2024.12		54.2	-
LLaVA-OV-7B 2024.12		52.9	-
VideoChat-Flash-7B 2026.01		51.3	-
LLaVA-Mini-8B 2025.03		51.2	-
MiniCPM-V-4.5-8B 2026.01		49.6	-
Video-LLaVA-7B 2025.03		45.3	-
LLaMA-VID-7B 2025.03		38.5	-
Mobile-VideoGPT-1.5B 2025.03		36.7	-
Video-ChatGPT-3.8B 2025.03		36.2	-
Mobile-VideoGPT-0.5B 2025.03		31.4	-
LLaVA-OneVision-0.5B 2025.03		26.8	-