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

Video Question Answering on Video-MME

86.9Accuracy

Gemini-3.0-Flash

Updated 23d ago

Evaluation Results

Method	Links
Gemini-3.0-Flash 2026.01		86.9
Gemini-2.5-Pro 2026.01		85.2
Qwen3-VL-8B + GRPO 2026.01		76.78
Full Precision 2026.07		73.3
Qwen3-VL-8B (baseline) 2026.01		71.4
Full Precision 2026.07		70.5
Qwen2-VL-7B + GRPO 2026.01		69.12
Qwen2-VL-7B (baseline) 2026.01		67.5
InternVideo2.5-8B 2026.01		65.4
GPTQ 2026.07		65.11
SAB-LVLM 2026.07		64.93
HINT InternVL3-8B 2026.03		64.6
InternVL3-8B 2026.03		64.2
LLaVA-Video-7B 2026.01		63.3
Qwen2-VL-2B + GRPO 2026.01		62.11
Full Precision 2026.07		61.96
GPTQ 2026.07		61.67
ARB-LLM 2026.07		61.15
GPTQ 2026.07		59.93
GPT-4V 2026.01		59.9
SAB-LVLM 2026.07		59.15
BiLLM 2026.07		58.41
ARB-LLM 2026.07		57.7
PB-LLM 2026.07		54.37
Qwen2-VL-2B (baseline) 2026.01		50.1
SAB-LVLM 2026.07		48.11
PB-LLM 2026.07		47.89
ARB-LLM 2026.07		47.14
BiLLM 2026.07		47.04
Video-LLaVA-7B 2026.01		45.3
BiLLM 2026.07		26.56
PB-LLM 2026.07		9.89