Multi-modal Visual Reasoning on VKnowU benchmark leaderboard · SOTA2 Research

Multi-modal Visual Reasoning on VKnowU

93Overall AccuracyHuman Performance

Evaluation Results

Method	Links
Human Performance2025.11		93	97.5	93.5	96.8	95.4	95.8	87.9	88.1	89.2	93.6	90.6
Gemini-3.0-Pro2025.11		76.3	58	90	91.6	56	73.9	74.8	73.6	81.7	81.8	78.5
Gemini-3.0-Flash2025.11		76.1	60.5	84	94.2	53.4	73	73.7	77.7	85	80	78.8
InternVL3.5-241B-A28B2025.11		74.6	52.5	67.5	85.8	60.6	66.4	81.6	77.2	84.2	83.6	81.9
InternVL3.5-38B2025.11		72.7	49	61	82.6	59.1	62.7	77.9	76.1	82.5	85.4	81.4
InternVL3.5-38B-Think2025.11		71.9	49	66.5	79	54.9	62.2	80.5	76.1	80.8	82.1	80.3
InternVL3.5-30B-A3B2025.11		71.5	49.5	65	82.1	62.7	64.6	75.8	77.2	81.7	77.2	77.5
Gemini-2.5-Pro2025.11		71.1	55	82	88.4	55.4	70.1	71.6	73.1	75.8	70.3	72
Qwen2.5-VL-72B-Instruct2025.11		70.1	59	64.5	86.3	40.9	62.6	76.8	79.7	82.5	73.3	76.7
LLaVA-Video-72B2025.11		69.9	57	65.5	86	44.5	63.1	78	72.6	84.2	73.8	75.8
Qwen3-VL-8B-Instruct2025.11		69	51.5	57.5	83.7	57	62.2	76.8	75.6	80	72.3	75
Gemini-2.5-Flash2025.11		68.8	56	80	90	51.8	69.3	76.8	65.5	74.2	63.6	68.2
GLM-4.6V-Flash2025.11		68.4	56	70.5	82.1	57	66.3	76.8	70	80.8	64.1	70.3
MiMo-VL-7B-RL2025.11		68	56	66	80.5	49.2	62.8	78.4	75.6	80.8	65.4	72.5
GLM-4.1V-9B-Thinking2025.11		67.6	54	66.5	80.5	47	61.9	72.5	73.1	80.8	69.8	72.6
Qwen3-VL-4B-Instruct2025.11		67.6	52	53	84.2	54.9	60.8	76.8	73.6	79.2	70.3	73.6
InternVL3.5-8B-Think2025.11		67.6	50.5	63	80.5	42	58.9	76.3	74.1	78.3	74.1	75.1
LLaVA-OV-72B2025.11		66.9	56	60	83	41	59.9	76	71.6	83.3	69	73
InternVL3.5-8B2025.11		66.9	50.5	58	78.9	44.6	57.9	75.3	75.6	75.8	73.8	74.8
GPT-5-mini2025.11		66.7	56.5	77	87.9	52.3	68.3	69.5	60.9	75.8	62.3	65.3
MiniCPM-V 4.52025.11		65.9	50	58.5	83.5	39	57.6	78.5	77.2	76.7	72.5	73.1
GPT-4o2025.11		65.7	55	74	84.7	42.5	64	74.7	63.5	65	65.9	67.1
LLaVA-Video-7B2025.11		64.7	50.5	56.5	79	42	56.9	72.5	69.5	80	69.5	71.5
mPLUG-Owl3-7B2025.11		64.2	57	55	72.6	36.3	55.2	77.9	73.1	79.2	66.7	72.1
Qwen2.5-VL-7B-Instruct2025.11		64	48	53	82.6	35.2	54.5	79.5	75.1	77.5	65.6	72.2
LLaVA-OV-7B2025.11		63.3	52.5	56	83	38	57.2	73	72.1	75.8	62.5	68.6
MiniCPM-V 2.62025.11		62.3	53	51.5	76.5	36.5	54.3	74	66.5	75	66.5	69.2
Qwen2.5-VL-3B-Instruct2025.11		61.1	50	57	77.9	38.3	55.7	71.6	73.1	71.7	57.4	65.8
VideoLLaMA2-7B2025.11		55.7	46	53	51.1	40.9	47.8	68.4	64.5	57.5	60.5	62.7
Random Guess2025.11		37.4	50	50	50	31.9	45.5	50	25	25	25	30.3