List your company

Video Quality Understanding on Q-Bench-Video (dev)

76.78Yes-or-No AccGemini-3-Pro
55.761661.218366.67572.1317Jan 26, 2026

Evaluation results

MethodLinks
Gemini-3-ProRequires VQA-labeled data=false, Proprietary=true2026.0176.786147.876169.5659.496063.02
Q-Router (GPT-4o)Requires VQA-labeled data=true, Reasoning backbone=GPT-4o2026.017657.0143.3359.3165.3161.4650.2260.07
OmniVQA-ChatterRequires VQA-labeled data=true2026.0175.5159.7640.3762.0561.5863.4542.3859.08
VQA^2Requires VQA-labeled data=true2026.0173.8156.438.3360.756.6561.0938.1156.67
QualiRAGRequires VQA-labeled data=false2026.0173.1862.9544.9561.3470.7461.6948.1761.66
VQAThinkerRequires VQA-labeled data=true2026.0170.4256.5635.1555.9564.5661.6930.9155.7
GPT-5Requires VQA-labeled data=false, Proprietary=true2026.0168.7256.4943.2156.5868.9959.9137.4257.22
Qwen2.5-VL-7BRequires VQA-labeled data=false2026.0168.5457.0135.5654.9463.1257.2735.755.3
GLM-4.1V-9BRequires VQA-labeled data=false2026.0163.8547.9631.4549.2860.7750.5924.2549.15
Qwen3-VL-8BRequires VQA-labeled data=false2026.0162.4451.8138.9251.6862.8554.931.9452.11
LLaVA-OneVision-7BRequires VQA-labeled data=false2026.0162.1352.2338.5648.7461.5348.8144.5752.12
InternVL3.5-8BRequires VQA-labeled data=false2026.0161.9751.1335.3149.2865.1754.3329.1450.7
mPLUG-Owl3-7BRequires VQA-labeled data=false2026.0160.8256.5235.8451.3460.4654.2637.352.44
InternVL3-9BRequires VQA-labeled data=false2026.0156.5751.5830.446.3559.850.8331.5147.61