List your company
ResearchFormerly WizWand

Multimodal Understanding on POPE

0.906POPE ScoreInternVL2.5
0.784320.815910.84750.87909Oct 17, 2024Jan 23, 2025May 2, 2025Aug 9, 2025Nov 15, 2025Feb 22, 2026Jun 1, 2026
Updated 4mo ago

Evaluation results

MethodLinks
InternVL2.5LLM=InternLM2.5-7B, Token Type=2D-Continuous dynamic2026.050.906——
InternVL2.5LLM=InternLM2.5-7B2025.050.906——
X-Omni# Tokens=~1T, # Params.=10B / 20B, Architecture Category=Autoregressive meets Diffusion2025.090.893——
LLaVA-OVScale=1.5-8B2026.060.892——
InternVL3-2BFT-Data=-, Fine-tuned on identical training data=false2026.030.889——
LatentUMBasequantized setting=false2026.040.889——
ILLUMEType=Und. and Gen., LLM Params=7B2025.010.885——
ILLUMELLM=Vicuna-7B, Token Type=2D-Continuous, Res.=2242026.050.885——
UnisonStage=Uni. 2 Stage2025.120.883——
MOSS-Video-PreviewSFT Mode=real-time2026.060.8817——
JanusFlowParams=1.3B, Architecture category=Integrating Autoregressive and Diffusion within Transformers2026.030.88——
SpaceR-sft-7BFT-Data=151K, Fine-tuned on identical training data=false2026.030.88——
JanusFlow2026.040.88——
MOSS-Video-PreviewSFT Mode=offline2026.060.8788——
Qwen2.5-VL-7BFT-Data=-, Fine-tuned on identical training data=false2026.030.878——
Tokenflow XL2026.040.878——
Qwen2.5-VLScale=7B2026.060.8768——
Harmo2026.040.876——
Qwen2.5-VL-3BFT-Data=-, Fine-tuned on identical training data=false2026.030.875——
Janus-Pro-7BType=Und. and Gen., LLM Params=7B2025.010.874——
Janus-Pro-7BStage=Uni. 1 Stage2025.120.874——
Janus-Pro# Tokens=~300B, # Params.=7B, Architecture Category=Autoregressive w/ Semantic Encoder2025.090.874——
Janus-ProParams=7B, Architecture category=Unifying Understanding and Generation via Next-token Prediction2026.030.874——
Janus Pro2026.040.874——
Janus-ProLLM=DeepSeek-LLM-7B, Token Type=2D-Continuous, Res.=3842026.050.874——
JanusType=Und. and Gen., # LLM Params=1.3B, External pretrained diffusion model=false2024.100.87——
JanusType=Und. and Gen., LLM Params=1.5B2025.010.87——
JanusParams=1.5B, Architecture category=Unifying Understanding and Generation via Next-token Prediction2026.030.87——
Qwen2.5-VL 7BToken Budget=All 1296 Tokens, Reduction Percentage=0%2026.030.87—100
VG-LLMFT-Data=385k, Fine-tuned on identical training data=false2026.030.869——
TokenFlow-XLType=Und. and Gen., LLM Params=13B2025.010.868——
TokenFlow-13BStage=Uni. 1 Stage2025.120.868——
TokenFlowParams=13B, Architecture category=Unifying Understanding and Generation via Next-token Prediction2026.030.868——
Cambrain-S-3BFT-Data=10M, Fine-tuned on identical training data=false2026.030.868——
WinTokLLM=Qwen3-8B, Token Type=1D-Continuous, Res.=2562026.050.865——
LLaVA-NeXTLLM=Vicuna-7B2025.050.865——
WallarooParams=7B, Architecture category=Unifying Understanding and Generation via Next-token Prediction, VLMEvalKit=true2026.030.864——
Qwen2.5 VLParams=7B, Architecture category=Unifying Understanding and Generation via Next-token Prediction, VLMEvalKit=true2026.030.863——
Janus-Pro-1BType=Und. and Gen., LLM Params=1.5B2025.010.862——
Janus-Pro-1BStage=Uni. 1 Stage2025.120.862——
LLaVA-v1.5Type=Und. Only, # LLM Params=7B, External pretrained diffusion model=false2024.100.859——
LLaVA-v1.5Type=Und. Only, LLM Params=7B2025.010.859——
LLaVA-v1.5LLM=Vicuna-7B, Token Type=2D-Continuous, Res.=3362026.050.859——
VARGPTLLM=Vicuna-7B, Token Type=2D-Continuous, Res.=2562026.050.859——
VILA-UType=Und. and Gen., # LLM Params=7B, External pretrained diffusion model=false2024.100.858——
VILA-UType=Und. and Gen., LLM Params=7B2025.010.858——
VILA-UStage=Uni. 1 Stage2025.120.858——
VILA-UParams=7B, Architecture category=Unifying Understanding and Generation via Next-token Prediction2026.030.858——
VILA-U2026.040.858——
SpatialLadder-3BFT-Data=26K, Fine-tuned on identical training data=false2026.030.855——
LatentUMBasequantized setting=true2026.040.855——
GeoSenseFT-Data=940K, Fine-tuned on identical training data=false2026.030.854——
Emu3-ChatType=Und. Only, # LLM Params=8B, External pretrained diffusion model=false2024.100.852——
Emu3-ChatType=Und. Only, LLM Params=8B2025.010.852——
EMU3# Tokens=-, # Params.=8B, Architecture Category=Autoregressive w/o Semantic Encoder2025.090.852——
Qwen2.5-VL-3B*FT-Data=940k, Fine-tuned on identical training data=true2026.030.852——
EMU32026.040.852——
Emu3LLM=8B (from scratch), Token Type=2D-Discrete, Res.=5122026.050.852——
VQRAELLM=Vicuna-13B, Token Type=2D-Continuous, Res.=2562026.050.851——
LLaVA-PhiType=Und. Only, # LLM Params=2.7B, External pretrained diffusion model=false2024.100.85——
LLaVA-PhiType=Und. Only, LLM Params=2.7B2025.010.85——
TokenFlow-LLLM=Vicuna-13B, Token Type=2D-Discrete, Res.=2562026.050.85——
MobileVLMType=Und. Only, # LLM Params=2.7B, External pretrained diffusion model=false2024.100.849——
MobileVLMType=Und. Only, LLM Params=2.7B2025.010.849——
TokLIPLLM=Qwen2.5-7B-Instruct, Token Type=1D-Continuous, Res.=3842026.050.849——
ViLASR-7BFT-Data=73K, Fine-tuned on identical training data=false2026.030.848——
MobileVLM-V2Type=Und. Only, # LLM Params=2.7B, External pretrained diffusion model=false2024.100.847——
MobileVLM-V2Type=Und. Only, LLM Params=2.7B2025.010.847——
Uni-X# Tokens=240B, # Params.=3B / 4.5B, Architecture Category=Autoregressive w/o Semantic Encoder, Extended Image-Text Training (flag)=true2025.090.846——
MobileVLMType=Und. Only, # LLM Params=1.4B, External pretrained diffusion model=false2024.100.845——
MobileVLMType=Und. Only, LLM Params=1.4B2025.010.845——
Show-o# Tokens=~500B, # Params.=1.3B, Architecture Category=Autoregressive meets Diffusion, Semantic Alignment (flag)=true2025.090.845——
VARGPT-9BStage=Uni. 2 Stage2025.120.844——
MobileVLM-V2Type=Und. Only, # LLM Params=1.4B, External pretrained diffusion model=false2024.100.843——
MobileVLM-V2Type=Und. Only, LLM Params=1.4B2025.010.843——
LLaVA-v1.5-Phi-1.5Type=Und. Only, # LLM Params=1.3B, External pretrained diffusion model=false2024.100.841——
LLaVA-v1.5-Phi-1.5Type=Und. Only, LLM Params=1.3B2025.010.841——
SEED-XLLM=LLaMA2-13B, Token Type=2D-Continuous, Res.=4482026.050.841——
WinTokLLM=Qwen3-8B, Token Type=1D-Continuous, Res.=256, semantic tokens=642026.050.841——
D-DitType=Und. and Gen., LLM Params=2.0B2025.010.84——
VILA-U# Tokens=-, # Params.=7B, Architecture Category=Autoregressive w/ Semantic Encoder2025.090.839——
VILA-ULLM=LLaMA2-7B, Token Type=2D-Discrete, Res.=2562026.050.839——
Uni-X# Tokens=140B, # Params.=3B / 4.5B, Architecture Category=Autoregressive w/o Semantic Encoder2025.090.836——
SemHiTokLLM=Qwen2.5-7B-Instruct, Token Type=2D-Discrete, Res.=2562026.050.834——
Liquid# Tokens=-, # Params.=8B, Architecture Category=Autoregressive w/ Semantic Encoder, Semantic Alignment (flag)=true2025.090.832——
UniTokLLM=LLaMA2-7B, Token Type=2D-Discrete, Res.=2562026.050.832——
Slot-MLLM (7B)LLM=Vicuna-7B2025.050.83——
Slot-MLLM (14B)LLM=Qwen2.5-14B-Instruct2025.050.822——
PromPruneToken Budget=512, Reduction Percentage=60.5%2026.030.818—94.8
Liquid# Tokens=~90B, # Params.=7B, Architecture Category=Autoregressive w/o Semantic Encoder2025.090.811——
LiquidLLM=Gemma-7B, Token Type=2D-Discrete, Res.=5122026.050.811——
DivPruneToken Budget=512, Reduction Percentage=60.5%2026.030.809—94.2
PromPruneToken Budget=256, Reduction Percentage=80.2%2026.030.807—92.6
Show-oType=Und. and Gen., LLM Params=1.3B, Resolution=5122025.010.8——
Show-oStage=Uni. 1 Stage2025.120.8——
Show-o# Tokens=~500B, # Params.=1.3B, Architecture Category=Autoregressive meets Diffusion2025.090.8——
Show-oLLM=Phi-1.5-1.3B, Token Type=2D-Discrete, Res.=2562026.050.8——
Orthus2026.040.796——
InstructBLIPType=Und. Only, # LLM Params=13B, External pretrained diffusion model=false2024.100.789——
InstructBLIPType=Und. Only, LLM Params=13B2025.010.789——