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

Visual Question Answering on MMStar

82.96Accuracy

Gemini 3-Pro

Updated 1mo ago

Evaluation Results

Method	Links
Gemini 3-Pro 2026.02		82.96
GPT-5 2026.02		82.1
Qwen3-VL 32B 2026.01		79.4
Gemini 2.5-Pro 2026.02		77.5
Qwen3-VL 2026.02		76.88
Gemini-2.5 Flash 2026.01		76.5
ERNIE 5.0 2026.02		75.54
Qwen3-VL 30B-A3B 2026.01		75.5
Qwen3-VL 8B 2026.01		75.3
MMFineReason-8B 2026.01		75.2
GPT5 mini 2026.01		74.1
HoneyBee 8B 2026.01		73.3
MMFineReason-4B 2026.01		72.8
SwimBird 2026.02		71.2
Qwen2.5-VL-32B-Instruct 2026.02		70.3
MMR1 8B 2026.01		69.3
OMR 7B 2026.01		69
MMFineReason-2B 2026.01		67.7
SkiLa 2026.02		64.8
Qwen3-VL-8B-Instruct 2026.02		64.7
Gemini-Pro 2024.07		64.1
LLaVA-OneVision 2026.02		61.9
GPT-4o 2024.06		61.6
Qwen2.5-VL-7B-Instruct 2026.02		60.3
IXC-2.5-7B 2024.07		59.9
InternVL-Chat-v1.5 2024.06		57.1
InternVL 2024.07		57.1
GPT-4V 2024.07		57.1
InternLM-XComposer2 2024.06		56.2
LLaVA-NeXT (Yi-34B) 2024.06		51.6
GPT-4v 2024.06		49.7
Avg 2026.02		49.03
MaD-Mix 2026.02		48.79
Uniform 2026.02		48.18
Fused 2026.02		46.55
Prism Captioner-7B 2024.06		45.9
Prism Captioner-7B 2024.06		43.7
Prism Captioner-2B 2024.06		43.3
Prism Captioner-2B 2024.06		42
LLaVA-InternLM2-20B 2024.06		41.9
MRoPE 2026.03		41.67
MRoPE-I 2026.03		40.73
Emu2-Chat 2024.06		40.7
Yi-VL-34B 2024.06		40.5
DeepSeek-VL-7B 2024.06		40.5
LLaVA-NeXT (Vicuna-13B) 2024.06		40.4
Vanilla RoPE 2026.03		40.2
MiniCPM-V-2 2024.06		39.1
MRoPE 2026.03		38.8
ShareCaptioner 2024.06		38.7
Vanilla RoPE 2026.03		38.53
MRoPE-I 2026.03		38.47
LLaVA-NeXT (Mistral-7B) 2024.06		38.4
LLaVA-InternLM2-7B 2024.06		38.3
LLaVA-NeXT (Vicuna-7B) 2024.06		37.6
Fused 2026.02		36.14
Avg 2026.02		35.54
mPLUG-Owl2 2024.06		34.8
LLaVA-InternLM2-1.8B 2024.06		34.5
LLaVA-v1.5-13B 2024.06		34.3
MaD-Mix 2026.02		33.97
LLaVA-v1.5-7B 2024.06		33.1
Uniform 2026.02		30.43