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

Multi-choice Video Question Answering on VideoMME

75Accuracy (no subs)

Gemini-1.5-Pro

Updated 1mo ago

Evaluation Results

Method	Links
Gemini-1.5-Pro 2025.03		75	81.3	-
GPT-4o 2025.03		71.9	77.2	-
Gemini-1.5-Flash 2025.03		70.3	75	-
InternVL2.5-8B-ITG-32 2025.07		67.3	69.6	-
Qwen2.5-VL-7B 2025.07		65.1	71.6	-
InternVL2.5-8B 2025.07		64.2	66.9	-
LLaVAction-7B 2025.03		63.9	71.4	-
LLaVA-Video-7B 2025.03		63.3	69.7	-
LLaVA-Video-7B 2025.07		63.3	69.7	-
LongVU-7B 2025.07		60.6	-	-
VILA-40B 2025.07		60.1	61.1	-
GPT-4V 2025.03		59.9	63.3	-
mPLUG-Owl3 2025.03		59.3	68.1	-
LLaVA-OV-7B 2025.03		58.2	61.5	-
LLaVA-OV-7B 2025.07		58.2	61.5	-
mPLUG-Owl3-8B 2025.07		53.5	-	-
LongVA-7B 2025.03		52.6	54.3	-
LongVA-7B 2025.07		52.6	54.3	-
VideoLLaMA2-7B 2025.03		47.9	50.3	-
VideoLLaMA2-7B 2025.07		45.1	46.6	-
VideoChat2-7B 2025.03		42.3	54.6	-
Video-LLaVA 2026.02		-	-	39.9
IG-VLM 2026.02		-	-	39.7
SF-LLaVA-7B 2026.02		-	-	39.4
DYTO 2026.02		-	-	41.2
KTV-7B-sparse 2026.02		-	-	43.6
KTV-7B-normal 2026.02		-	-	43.7
KTV-7B-dense 2026.02		-	-	44
IG-VLM 2026.02		-	-	50.3
SF-LLaVA-7B 2026.02		-	-	51.9
DYTO 2026.02		-	-	53.4
KTV-34B-sparse 2026.02		-	-	52.2
KTV-34B-normal 2026.02		-	-	53
KTV-34B-dense 2026.02		-	-	53.2