List your company
ResearchFormerly WizWand

Vision-centric Evaluation on CV-Bench

0.8751AccuracyQwen3-VL-32B
0.3448040.4824770.620150.757823Jan 12, 2026Feb 7, 2026Mar 5, 2026Mar 31, 2026Apr 26, 2026May 22, 2026Jun 17, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Qwen3-VL-32B3-run average=true2026.060.8751
SmoothOp-7BData Scale=50K, Training Phases=1 (R)2026.010.864
InternVL3.5-38B3-run average=true2026.060.8596
Qwen3-VL-8B3-run average=true2026.060.8563
VST-7B-SFTData Scale=4.1M, Training Phases=1 (S)2026.010.855
GPT-53-run average=true2026.060.8546
SmoothOp-3BData Scale=50K, Training Phases=1 (R)2026.010.849
VST-3B-SFTData Scale=4.1M, Training Phases=1 (S)2026.010.844
InternVL3.5-8B3-run average=true2026.060.8199
ViLaSR-7BData Scale=81K, Training Phases=2 (S+R)2026.010.818
ThinkMorph3-run average=true, Latency (s)=142.82026.060.8086
SPAR-8BData Scale=7M, Training Phases=2 (S)2026.010.807
Visual-OPSD3-run average=true, Latency (s)=10.02026.060.8064
Visual-OPSD-Noise3-run average=true, Latency (s)=14.92026.060.7909
Text-only SFT3-run average=true, Latency (s)=28.52026.060.7737
BAGEL-7B3-run average=true2026.060.7603
GPT-4oData Scale=-, Training Phases=-2026.010.76
GPT-4o3-run average=true2026.060.7561
Qwen2.5-VL-7BData Scale=-, Training Phases=-2026.010.754
SpaceR-7BData Scale=151K, Training Phases=2 (S+R)2026.010.748
SpatialLadder-3BData Scale=26K, Training Phases=3 (S+R)2026.010.737
Qwen2.5-VL-3BData Scale=-, Training Phases=-2026.010.714
Janus-Pro-7B3-run average=true2026.060.6783
LLava-OneVision-7BData Scale=-, Training Phases=-2026.010.619
SPHINX-13B2026.010.613
CARPE2026.010.588
CARPE-MoE2026.010.588
Gemini-2.5-ProData Scale=-, Training Phases=-2026.010.535
LEVI2026.010.474
WiSE-FT2026.010.461
LLaVA1.5-7B2026.010.46
InternVL3-8BData Scale=-, Training Phases=-2026.010.421
LLaVA1.5-7BFine-tuned=ImageNet2026.010.379
Chameleon-7B3-run average=true2026.060.3652