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

Video Reasoning on Video-Holmes (Accuracy)

60.5Accuracy

VideoChat-M1

Updated 8d ago

Evaluation Results

Method	Links
VideoChat-M1 2025.11		60.5
OmniJigsaw (CMM) 2026.04		52.53
OmniJigsaw (SMS) 2026.04		52.26
VideoJigsaw 2026.04		51.99
Qwen3-Omni-30B 2026.04		50.84
OmniJigsaw (JMI) 2026.04		50.24
OmniJigsaw (CMM) 2026.04		48.29
OmniJigsaw (SMS) 2026.04		47.96
VideoJigsaw 2026.04		47.9
OmniJigsaw (JMI) 2026.04		47.47
Qwen3-Omni-30B 2026.04		46.6
Gemini 1.5 Pro 2025.11		45.7
HumanOmniV2 2026.04		42.9
Qwen3-VL-8B-Thinking + MVP 2026.01		42.6
Video-R1 2026.04		42.13
GPT-4o 2025.11		42
Qwen3-VL-8B-Thinking 2026.01		41.8
Omni-R1 2026.04		40.72
Qwen3-VL-8B-Thinking + MVP 2026.01		40.1
InternVL3.5-8B + MVP 2026.01		39.4
InternVL3.5-8B 2026.01		39.3
HumanOmniV2 2026.04		38.87
InternVL3.5-8B + MVP 2026.01		38.8
Omni-R1 2026.04		38.76
Qwen3-VL-8B-Thinking 2026.01		38.5
InternVL3.5-8B 2026.01		37.7
Jigsaw-7B 2026.01		37.6
Jigsaw-7B 2026.01		36.9
Qwen2.5-VL-7B-Instruct + MVP 2026.01		36.7
VideoChat-R1-7B 2025.11		36.5
Qwen2.5-VL-7B-Instruct + MVP 2026.01		35.9
Qwen2.5-VL-7B-Instruct 2026.01		35
Qwen2.5-VL-7B-Instruct 2026.01		33.4
VideoChat-R1-7B (Alternative) 2025.11		33
InternVL-3-8B 2025.11		32.3
Qwen2-VL-7B 2025.11		27.8
InternVL-2.5-8B 2025.11		23.6