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

Video Understanding on MMVU

78.2Accuracy

Seed2.0 Pro

Updated 18d ago

Evaluation Results

Method
Seed2.0 Pro 2026.06	78.2	-	-	-	-
Gemini-3-Flash 2026.06	77.9	-	-	-	-
Gemini-3-Pro 2026.06	76.3	-	-	-	-
GPT-4o 2026.05	75.4	-	-	-	-
Seed2.0 Lite 2026.06	75	-	-	-	-
Qwen3-VL-8B + ADPO 2026.05	73.1	-	-	-	-
Seed1.8 2026.06	73.1	-	-	-	-
Qwen3-VL-8B + DAPO 2026.05	72	-	-	-	-
VideoSearcher-8B 2026.07	70.4	-	-	-	-
Qwen3-VL-8B + GRPO 2026.05	70.2	-	-	-	-
Qwen3-VL-8B 2026.05	69.7	-	-	-	-
Qwen3-VL-8B-Instruct 2026.07	69.28	-	-	-	-
Seed2.0 Mini 2026.06	69	-	-	-	-
VideoSearcher-4B 2026.07	68.64	-	-	-	-
ParaVT-8B 2026.05	68.6	-	-	-	-
VideoRFT-7B 2026.05	68.5	-	-	-	-
VideoRFT 2026.07	68.5	-	-	-	-
Qwen3-VL-8B 2026.05	68	-	-	-	-
EvoVid 2026.05	67	-	-	-	-
GPT-4o 2026.05	66.7	-	-	-	-
Gemini-2.0-Flash 2026.05	66.5	-	-	-	-
Video-RTS-7B 2026.05	66.4	-	-	-	-
EvoVid 2026.05	66.2	-	-	-	-
ViSS-R1-7B 2026.05	66.1	-	-	-	-
VidGround 2026.04	65.8	-	-	-	-
Gemini 1.5 Pro 2026.05	65.8	-	-	-	-
VidGround 2026.04	65.6	-	-	-	-
Qwen2.5-VL-7B + SAGE 2026.05	65.4	-	-	-	-
Qwen2.5-VL-7B 2026.05	65.4	-	-	-	-
VideoCoM 2026.07	65.4	-	-	-	-
Qwen2.5-VL-7B + SFT + SAGE 2026.05	65.3	-	-	-	-
Video-RTS 2026.04	65	-	-	-	-
VideoChat-R1-7B 2026.05	65	-	-	-	-
Base Model 2026.05	64.8	-	-	-	-
Video-Thinker-7B 2026.05	64.5	-	-	-	-
VidGround 2026.04	64.2	-	-	-	-
TW-GRPO 2026.04	64.2	-	-	-	-
Video-R1-7B 2026.05	64.2	-	-	-	-
Conan-7B 2026.05	64	-	-	-	-
Video-RTS 2026.04	63.9	-	-	-	-
EvoVid 2026.05	63.8	-	-	-	-
Video-R1 2026.07	63.8	-	-	-	-
EvoVid 2026.05	63.7	-	-	-	-
Time-R1-7B 2026.05	63.4	-	-	-	-
LongVT-RFT-7B 2026.05	63.4	-	-	-	-
Qwen3-VL-4B-Instruct 2026.07	63.36	-	-	-	-
TW-GRPO 2026.04	63.1	-	-	-	-
Qwen2.5-VL-7B 2026.04	62.6	-	-	-	-
EvoVid 2026.05	62.6	-	-	-	-
EvoVid 2026.05	62.4	-	-	-	-
EvoVid 2026.05	62.4	-	-	-	-
Qwen2.5-VL-7B 2026.04	62.3	-	-	-	-
VideoChat-R1.5 2026.07	62.08	-	-	-	-
TW-GRPO 2026.04	61.8	-	-	-	-
Video-RTS 2026.04	61.8	-	-	-	-
VideoZoomer-7B 2026.05	61.6	-	-	-	-
LongVILA-R1-7B 2026.04	61.5	-	-	-	-
Video-R1-7B 2026.05	61.3	-	-	-	-
EvoVid 2026.05	61.1	-	-	-	-
EvoVid 2026.05	61	-	-	-	-
EvoVid 2026.05	60.6	-	-	-	-
Qwen2.5-VL-7B 2026.04	60.5	-	-	-	-
Qwen2.5-VL-7B + SFT 2026.05	60.5	-	-	-	-
EvoVid 2026.05	60.5	-	-	-	-
EvoVid 2026.05	60.5	-	-	-	-
Base Model 2026.05	60.3	-	-	-	-
ReWatch-R1-7B 2026.05	59.8	-	-	-	-
Qwen2.5-VL-7B 2026.05	59.2	-	-	-	-
Base Model 2026.05	59.2	-	-	-	-
LongVILA-R1-7B 2026.04	59.1	-	-	-	-
LongVILA-R1-7B 2026.04	58.8	-	-	-	-
Qwen3-VL 2026.03	58.7	-	-	-	-
Video-R1 2026.04	56.2	-	-	-	-
SAGE-7B 2026.05	55.7	-	-	-	-
Qwen2.5-VL-7B-SFT 2026.04	55.4	-	-	-	-
EvoVid 2026.05	54.6	-	-	-	-
Video-R1 2026.04	54.5	-	-	-	-
Penguin-VL 2026.03	53.9	-	-	-	-
EvoVid 2026.05	53.3	-	-	-	-
Video-R1 2026.04	53.2	-	-	-	-
EvoVid 2026.05	53.1	-	-	-	-
EvoVid 2026.05	51.7	-	-	-	-
InternVL-3.5 2026.03	51.5	-	-	-	-
Qwen2.5-VL-7B-SFT 2026.04	51.3	-	-	-	-
OpenAI GPT-5 nano 2026.03	51	-	-	-	-
Qwen2.5-VL-7B-SFT 2026.04	51	-	-	-	-
Human 2026.06	49.7	-	-	-	-
LLaVA-OneVision-7B 2026.05	49.2	-	-	-	-
Base Model 2026.05	48.5	-	-	-	-
VideoLLaMA2 2026.05	44.8	-	-	-	-
VideoRFT-7B 2026.05	42.7	-	-	-	-
Video-LLaVA 2026.01	-	29	28.2	-	-
Video-LLaVA + TCD 2026.01	-	28.7	28.6	-	-
Video-LLaVA + MotionCD 2026.01	-	27.6	27.2	-	-
Video-LLaVA + Dino-Heal 2026.01	-	28.7	29.2	-	-
SSCD 2026.01	-	32	28.5	-	-
LLaVA-NeXT 2026.01	-	31.3	30.5	-	-
LLaVA-NeXT + TCD 2026.01	-	25.6	29.8	-	-
LLaVA-NeXT + MotionCD 2026.01	-	26.3	25.1	-	-
LLaVA-NeXT + Dino-Heal 2026.01	-	30.9	30.7	-	-

Showing 100 of 116 rows