List your company
ResearchFormerly WizWand

Multimodal Understanding on RealWorldQA

78RWQA ScoreInternVL3-78B
48.04855.82463.671.376Dec 4, 2024Feb 26, 2025May 21, 2025Aug 13, 2025Nov 5, 2025Jan 28, 2026Apr 23, 2026
Updated 6mo ago

Evaluation results

MethodLinks
InternVL3-78Bnumber of parameters=78B2026.0278
OmniInference scheme=no-thinking2026.0476
InternVL3-38Bnumber of parameters=38B2026.0275.6
Qwen3-VL-30B-A3B-InstructInference scheme=no-thinking2026.0473.7
InternVL3.5-30B-A3BInference scheme=no-thinking2026.0472.3
Innovator-VLvariant=8B-Instruct, parameters=8B2026.0171.5
Innovator-VLvariant=8B-Thinking, parameters=8B2026.0170.98
TPRU-32Bnumber of parameters=32B2026.0269.8
TPRU-7Bnumber of parameters=7B2026.0269.54
MiniCPM-Vversion=4.5, parameters=8B2026.0169.28
Qwen3-VLparameters=8B2026.0169.02
Qwen2.5-VLParams=7B2026.0469
Qwen3-VLParams=8B2026.0469
Qwen2.5-VL-32Bnumber of parameters=32B2026.0268.89
Qwen2.5-VL-7Bnumber of parameters=7B2026.0268.1
InternVL3Params=8B2026.0467.8
LLaVA-OVversion=1.5, parameters=8B2026.0167.71
MiMo-VLtraining=7B-SFT, parameters=7B2026.0167.19
TPRU-3Bnumber of parameters=3B2026.0266.54
MiMo-VLtraining=7B-RL, parameters=7B2026.0166.01
Qwen2.5-VL-3Bnumber of parameters=3B2026.0265.36
InternVL3.5parameters=8B2026.0164.71
Intern-S1variant=mini, parameters=9B2026.0163.66
InternVL3.5Params=8B2026.0463
GPT-5 Nano2026.0460.7
EMU3Backbone=8B from scratch, Resolution=512, Visual Input Type=Discrete2024.1257.4
TokenFlow-XLBackbone=Qwen-2.5-14B, Resolution=384, Visual Input Type=Discrete2024.1256.6
LLaVA-1.5Backbone=Vicuna-13B, Resolution=336, Visual Input Type=Continuous2024.1255.3
ShareGPT4VBackbone=Vicuna-7B, Resolution=336, Visual Input Type=Continuous2024.1254.9
TokenFlow-XLBackbone=Vicuna-13B, Resolution=384, Visual Input Type=Discrete2024.1253.7
UniDriveVLAParams=8B2026.0449.9
TokenFlow-BBackbone=Vicuna-13B, Resolution=224, Visual Input Type=Discrete2024.1249.4
TokenFlow-LBackbone=Vicuna-13B, Resolution=256, Visual Input Type=Discrete2024.1249.2