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

Multi-modal Understanding on MMBench (mean accuracy)

86.3Mean Accuracy

Oryx-1.5

Updated 1mo ago

Evaluation Results

Method	Links
Oryx-1.5 2024.09		86.3
LLaVA-OneVision 2024.09		85.6
Oryx 2024.09		84.5
Latent Denoising 2026.04		84.5
Qwen3-VL-4B 2026.04		83.58
Baseline 2026.04		83.1
VILA-1.5 2024.09		82.4
Cambrian-1 2024.09		81.4
Oryx 2024.09		81.4
Oryx-1.5 2024.09		81.3
LLaVA-OneVision-7B 2026.04		80.57
LLaVA-NeXT 2024.09		79.3
InfiniteVL-4B 2026.04		79
SpB2.0-VL-5B 2026.04		78.39
Bunny-LLama3 2024.09		77.2
Qwen2.5-VL-3B 2026.04		76.88
Idefics2 2024.09		76.7
Cambrian-1 2024.09		75.9
VILA-1.5 2024.09		75.3
InternVL2-4B 2026.04		73.6
Deepseek-VL 2024.09		73.2
Monkey 2024.09		72.4
LLaVA-NeXT 2024.09		72.1
VITA 2024.09		71.8
MaTVLM-3B 2026.04		69.4
Latent Denoising 2026.04		68.7
Vanilla 2026.06		68.5
PriorTR 2026.06		68.3
PDrop 2026.06		67.5
SmolVLM2-2.2B 2026.04		67.48
SparseVLM 2026.06		67.4
PriorTR 2026.06		67.3
VisPruner 2026.06		66.8
CAMD 2026.03		66.6
VisPruner 2026.06		66.5
FarSight 2026.03		66
SparseVLM 2026.06		65.8
PriorTR 2026.06		65.8
Latent Denoising 2026.04		65.7
PDrop 2026.06		65
LLaVA-1.5-7B 2025.08		64.7
EVIAN 2026.04		64.63
CGD 2026.03		64.5
Baseline 2026.04		64.5
OPERA 2026.03		64.4
LLaVA-1.5 2026.03		64.3
Baseline 2026.04		64
VCD 2026.03		63.9
HiPrune++ 2025.08		63.5
VisPruner 2026.06		63.2
SCALE 2026.04		63.18
BLIP-2 2026.04		63.17
ICD 2026.03		63.1
CAMD 2026.03		63.1
FarSight 2026.03		62.8
HiPrune 2025.08		62.8
HiPrune++ 2025.08		62.3
HiPrune 2025.08		62.2
BLIP 2026.04		61.83
SparseVLM 2026.06		61.3
Video-LLaVA 2026.03		60.9
CAMD 2026.03		60.6
HiPrune++ 2025.08		60.3
ALBEF 2026.04		60.03
FarSight 2026.03		59.8
Full Data 2026.04		59.53
HiPrune 2025.08		59.5
PDrop 2026.06		59.2
Qwen2.5-VL 2026.04		57.96
FastV 2026.06		57.9
CLIPScore 2026.04		57.46
Chat-UniVi 2026.03		56.3
Cobra-3B 2026.04		55.9
FastV 2026.06		54
Random 2026.04		53.53
FastV 2026.06		50.9
CAMD 2026.03		46.8
FarSight 2026.03		46.5
InstructBLIP 2026.03		43.4