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

Multimodal Understanding on MMBench (dev)

93.2MMB Score

Gemini 3.1 Pro

Updated 1mo ago

Evaluation Results

Method
Gemini 3.1 Pro 2026.06	93.2	-	-
Vanilla 2026.05	88.5	-	100
VSCAN 2026.05	88.5	-	94.4
G2TR 2026.05	88.5	-	99
IVC-Prune 2026.05	88.4	-	98.6
Qwen3-VL 32B 2026.06	87.8	-	-
Cosmos 3 Super 2026.06	87.4	-	-
Gemma-4 31B 2026.06	86.7	-	-
Cosmos-Reason2 32B 2026.06	86.3	-	-
W-FastV 2026.05	85.9	-	95.6
FastV 2026.05	85.7	-	95.8
RynnBrain 8B 2026.06	85.5	-	-
Qwen3-VL 8B 2026.06	85.3	-	-
Cosmos 3 Nano 2026.06	85.1	-	-
BAGAL 2025.03	85	-	-
Qwen3-VL 2026.02	84	-	-
PDrop 2026.05	84	-	94.1
SigLIP2 \| Qwen-3-8B 2026.02	82.5	-	-
Cosmos-Reason2 8B 2026.06	82.2	-	-
Qwen2.5-VL 2026.02	81.9	-	-
RynnBrain 2B 2026.06	81.9	-	-
MiMo-Embodied 7B 2026.06	81.4	-	-
SigLIP2 \| Qwen-3-4B 2026.02	81	-	-
Vanilla 2026.05	80.5	-	100
DualToken-7B 2025.03	80	-	-
LLaVA-NeXT 2025.03	79.3	-	-
Show-o2 2025.03	79.3	-	-
G2TR 2026.05	78.8	-	94.9
W-FastV 2026.05	78.6	-	92.3
IVC-Prune 2026.05	78.4	-	93.7
SigLIP \| Gemma3-4B 2026.02	78.3	-	-
VSCAN 2026.05	78.1	-	92.2
FastV 2026.05	77.9	-	92.9
PDrop 2026.05	77.7	-	92
SigLIP2 \| Lilium-8B 2026.02	77.4	-	-
Qwen3-VL 2B 2026.06	77.2	-	-
SigLIP2 \| e-Llama3.1-8B 2026.02	76.9	-	-
Cosmos 3 Edge 2026.06	76.6	-	-
LLaVA-OV 2026.02	76.4	-	-
DualToken-3B 2025.03	76.2	-	-
SigLIP2 \| Llama-3.1-8B 2026.02	75.8	-	-
SigLIP2 \| Lilium-4B 2026.02	75.5	-	-
DualToken-7B 2025.03	74.9	-	-
Cosmos-Reason2 2B 2026.06	73.6	-	-
MUSE-VL 2025.03	72.1	-	-
Qwen2.5ViT \| e-Llama3.1-8B 2026.02	71.7	-	-
UniToken 2025.03	71.1	-	-
DualToken-3B 2025.03	70.9	-	-
InternVL3.5-1B 2025.11	70.79	-	-
Janus 2025.03	69.4	-	-
LFM2-VL-1.6B 2025.11	69.16	-	-
VILA 2025.03	68.9	-	-
TokenFlow 2025.03	68.9	-	-
ShareGPT4V 2025.03	68.8	-	-
Gemma3 2026.02	67.9	-	-
LLaVA-NeXT 2025.03	67.4	-	-
Show-o2 2025.03	67.4	-	-
Gemma-4 E4B 2026.06	66.2	-	-
SigLIP2 \| Lilium-1B 2026.02	64.7	-	-
LLaVA-1.5 2025.03	64.3	-	-
UniTok 2025.03	61.1	-	-
Gemma-4 E2B 2026.06	60.2	-	-
LLaVA-Phi 2025.03	59.8	-	-
Emu3 2025.03	58.5	-	-
LFM2-VL-450M 2025.11	55.76	-	-
SmolVLM2-500M 2025.11	51.29	-	-
InstructBLIP 2025.03	36	-	-
Chameleon 2025.03	31.1	-	-
InfiMM-HD 2024.03	0.716	-	-
LLaVA 1.5 2024.03	0.677	-	-
Sphinx-2K 2024.03	0.659	-	-
Qwen-VL-Chat 2024.03	0.606	-	-
Shikra 2024.03	0.588	-	-
GPT-4V-1106 2024.02	-	75.1	-
LVIS-INSTRUCT4V-13B 2024.02	-	68	-
ShareGPT4V-13B 2024.02	-	68.5	-
VILA1.5-13B 2024.02	-	74.9	-
Mini-Gemini-HD-13B 2024.02	-	68.6	-
LLaVA-v1.5-13B (LORA) 2024.02	-	68.5	-
LLaVA-NeXT-13B 2024.02	-	70	-
InstructBLIP-Vicuna-7B 2024.02	-	36	-
Shikra-7B 2024.02	-	58.8	-
mPLUG-Owl2 2024.02	-	66.5	-
SPHINX-Intern2 2024.02	-	57.9	-
Yi-VL-6B 2024.02	-	68.6	-
DeepSeek-VL-7B 2024.02	-	73.2	-
LLaVA-v1.5-7B (LORA) 2024.02	-	66.1	-
LLaVA-NeXT-7B 2024.02	-	67.4	-
MM1-7B-Chat 2024.02	-	72.3	-
Idefics2 2024.02	-	76.7	-
LLaVA-NeXT-Llama-3-8B 2024.02	-	72.1	-
Bunny-8B 2024.02	-	77.2	-
MobileVLM-V2-3B 2024.02	-	63.2	-
VILA1.5-3B 2024.02	-	63.4	-
MiniCPM-V 2.0 2024.02	-	69.6	-
MM1-3B-Chat 2024.02	-	67.8	-
Imp-v1.5-4B-Phi3 2024.02	-	73.3	-
Bunny-4B 2024.02	-	74.2	-
MobileVLM 2024.11	-	59.6	-
MobileVLM-V2 2024.11	-	63.2	-

Showing 100 of 131 rows