List your company

Visual Spatial Instruction on VSI-Bench

40.5Relative Direction (Hard)LLaVA-OneVision-1.5-4B + SpatialSV
22.92427.48732.0536.613Jun 18, 2026

Evaluation results

MethodLinks
LLaVA-OneVision-1.5-4B + SpatialSVSupervision=SpatialSV, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.0640.535.252.54225.436.137.1
LLaVA-OneVision-1.5-4B + DistillationSupervision=Feature Distillation, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.064035.547.93924.63335.3
LLaVA-OneVision-1.5-4B + TextSupervision=Textual, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.0638.934.148.938.625.130.934.9
LLaVA-OneVision-1.5-8B + SpatialSVSupervision=SpatialSV, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.0638.336.549.342.129.640.238.1
LLaVA-OneVision-1.5-4BSupervision=None, Model Family=LLaVA-OneVision-1.5, Model Size=4B2026.0637.335.546.13724.329.933.9
LLaVA-OneVision-1.5-8B + DistillationSupervision=Feature Distillation, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.0636.535.550.238.328.33235.7
Qwen2.5-VL-3B + DistillationSupervision=Feature Distillation, Model Family=Qwen2.5-VL, Model Size=3B2026.0635.934.130.932.122.83230.6
Qwen2.5-VL-3B + TextSupervision=Textual, Model Family=Qwen2.5-VL, Model Size=3B2026.0635.735.726.331.123.629.930.1
Qwen2.5-VL-3B + SpatialSVSupervision=SpatialSV, Model Family=Qwen2.5-VL, Model Size=3B2026.0634.637.834.633.523.337.132.2
Qwen2.5-VL-3BSupervision=None, Model Family=Qwen2.5-VL, Model Size=3B2026.0634.333.32931.722.727.829.6
LLaVA-OneVision-1.5-8B + TextSupervision=Textual, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.0634.135.552.137.329.930.435.5
InternVL3-2B + SpatialSVSupervision=SpatialSV, Model Family=InternVL3, Model Size=2B2026.0631.435.247.935.128.841.234.6
InternVL3-1BSupervision=None, Model Family=InternVL3, Model Size=1B2026.0630.833.94726.624.633.530.2
InternVL3-8B + SpatialSVSupervision=SpatialSV, Model Family=InternVL3, Model Size=8B2026.0630.637.350.736.233.242.836.6
InternVL3-1B + SpatialSVSupervision=SpatialSV, Model Family=InternVL3, Model Size=1B2026.063034.750.228.224.938.731.4
InternVL3-1B + DistillationSupervision=Feature Distillation, Model Family=InternVL3, Model Size=1B2026.0629.835.243.828.225.13230.4
Qwen2.5-VL-7B + SpatialSVSupervision=SpatialSV, Model Family=Qwen2.5-VL, Model Size=7B2026.0629.535.248.437.230.142.835.4
LLaVA-NeXT-Video-7B + DistillationSupervision=Feature Distillation, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.0629.533.648.437.529.534.534.4
InternVL3-2B + DistillationSupervision=Feature Distillation, Model Family=InternVL3, Model Size=2B2026.0629.533.347332831.432.4
LLaVA-NeXT-Video-7B + SpatialSVSupervision=SpatialSV, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.0629.233.952.537.830.444.935.9
Qwen2.5-VL-7B + DistillationSupervision=Feature Distillation, Model Family=Qwen2.5-VL, Model Size=7B2026.062932.550.234.829.137.633.7
InternVL3-1B + TextSupervision=Textual, Model Family=InternVL3, Model Size=1B2026.062934.744.727.325.430.930
InternVL3-2B + TextSupervision=Textual, Model Family=InternVL3, Model Size=2B2026.062934.145.63230.129.932.4
LLaVA-OneVision-1.5-8BSupervision=None, Model Family=LLaVA-OneVision-1.5, Model Size=8B2026.0628.734.948.936.128.828.933.5
Qwen2.5-VL-7BSupervision=None, Model Family=Qwen2.5-VL, Model Size=7B2026.0628.424.945.231.829.531.430.8
InternVL3-8B + DistillationSupervision=Feature Distillation, Model Family=InternVL3, Model Size=8B2026.0628.43647.535.531.73434.5
LLaVA-NeXT-Video-7B + TextSupervision=Textual, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.0627.935.745.635.529.832.533.6
Qwen2.5-VL-7B + TextSupervision=Textual, Model Family=Qwen2.5-VL, Model Size=7B2026.0627.431.251.231.430.134.532.4
LLaVA-NeXT-Video-7BSupervision=None, Model Family=LLaVA-NeXT-Video, Model Size=7B2026.0627.132.550.735.927.531.432.9
Chance Level (Random)Type=Baseline2026.0626.525.724.924.123.328.427.7
InternVL3-2BSupervision=None, Model Family=InternVL3, Model Size=2B2026.0626.531.548.430.124.43430.3
InternVL3-8B + TextSupervision=Textual, Model Family=InternVL3, Model Size=8B2026.0625.733.650.234.53229.933.5
Chance Level (Frequency)Type=Baseline2026.0625.233.650.225.125.229.429
InternVL3-8BSupervision=None, Model Family=InternVL3, Model Size=8B2026.0623.634.146.134.132.932.533.1