Multimodal Video Question Answering on LiViBench 1.0 (test) benchmark leaderboard · SOTA2 Research

Multimodal Video Question Answering on LiViBench 1.0 (test)

66.2Overall ScoreSeed1.5-VL

Evaluation Results

Method	Links
Seed1.5-VL2026.01		66.2	70.9	71.4	68.8	70.7	59.1
Doubao-Seed-1.62026.01		64.9	72.9	73.2	60.2	68.4	56.8
InternVL3-78B2026.01		64.4	72	69.8	65.8	69.3	56.3
LiVi-LLM-7B2026.01		64.4	70.1	68.7	62.8	63.6	60.9
InternVL3-38B2026.01		64.1	70.9	72.6	66.6	68.3	54.5
InternVL3-14B2026.01		62.7	71.7	68.9	65.3	67	53.7
Qwen2.5-VL-72B2026.01		62.3	73.4	72.4	61.9	64.6	52
Qwen2.5-Omni-7B2026.01		60.3	68.1	68.5	59.4	60.7	53.1
LLaVA-Video-72B2026.01		60	65.3	70.2	63.6	63.8	49.8
InternVL3-9B2026.01		60	67.3	67.8	59.4	63	51.7
InternVL3-8B2026.01		59.8	68.4	66.9	58.5	63	51.7
Qwen2.5-VL-32B2026.01		59.4	73.1	69.1	57.6	61.3	49.1
Qwen2.5-VL-7B2026.01		58.3	65.5	69.2	57.6	59.7	49.1
InternVL2.5-8B2026.01		56.6	68.1	62.9	58.5	59.9	47.5
GPT-4o2026.01		56.3	67	66.5	57.6	55.2	47.4
Gemini 2.5 Pro2026.01		56.1	65	68.4	58.1	51.3	48.2
MiniCPM-o-262026.01		56	65.3	62.1	61.9	59.9	46.5
Keye-VL-8B-Preview2026.01		55.2	68.7	65.4	51.7	52.6	47.2
Video-LLaMA3-8B2026.01		54.1	60	62.1	61.9	58.8	43.7
NVILA-8B-Video2026.01		53.3	61.1	56.2	58.1	56.8	46.6
Gemini 2.5 Flash2026.01		53	63.6	62.9	56.4	51.5	43.9
LLaVA-Video-7B2026.01		52.6	60	59	55.1	58.3	43.5
MiniCPM-v-262026.01		50.5	58.9	59.9	58.5	54.2	39.1
InternVL2-8B2026.01		49.8	59.7	59.2	53.8	53.6	38.7
LLaVA-NeXT-Video2026.01		37.5	40.7	40.5	44.8	32.3	36.1