List your company
ResearchFormerly WizWand

Computer Vision Reasoning on CV-Bench

85.7AccuracyGAMSI_S1+S2
41.81253.20664.675.994May 8, 2026May 10, 2026May 13, 2026May 16, 2026May 19, 2026May 22, 2026May 25, 2026
Updated 5mo ago

Evaluation results

MethodLinks
GAMSI_S1+S2SI Dataset=950K2026.0585.7
VST-7B-SFTSI Dataset=4M2026.0585.5
GazeVLM (Ours)Gaze Bias=Enabled, Base Model=Qwen3-VL-4B2026.0585.4
VST-3B-SFTSI Dataset=4M2026.0584.6
GazeVLM (w/o gaze bias)Gaze Bias=None, Base Model=Qwen3-VL-4B2026.0584.5
SenseNova-SI-InternVL3-8BSI Dataset=800K2026.0584
GAMSI_S1SI Dataset=800K2026.0583.9
Qwen3-VL-4BTraining Paradigm=Vanilla Open-source, Parameters=4B2026.0583.4
SenseNova-SI-Qwen3-VL-8BSI Dataset=8M2026.0583.4
Qwen3-VL-8B-Instruct2026.0582.4
Qwen3.5-4B-InstTraining Paradigm=Vanilla Open-source, Parameters=4B2026.0581.1
Cambrian-S-7BSI Dataset=590K2026.0576.9
Qwen2.5-VL-7B-Instruct2026.0575.3
Cambrian-S-3BSI Dataset=590K2026.0575.2
SpaceR-7BSI Dataset=151K2026.0574.9
Qwen2.5-VL-7BTraining Paradigm=Vanilla Open-source, Parameters=7B2026.0573.7
SpatialLadder-3BSI Dataset=260K2026.0573.6
InternVL3-8B2026.0568.4
InternVL3-2B2026.0568.1
Qwen3-VL-2B-Instruct2026.0566.3
GPT-4oTraining Paradigm=Closed-source2026.0566
Gemini2.5-Flash-LiteTraining Paradigm=Closed-source2026.0564.9
Qwen2.5-VL-3B-Instruct2026.0563
Bagel-7B-MoT2026.0556.7
Spatial-MLLM-4BSI Dataset=135K2026.0554.7
ViLaSR-7BSI Dataset=81K2026.0543.5