List your company

Fine-grained Video Spatial Reasoning on VSI-Bench

60.6Avg ScoreVST-7B-SFT
31.27238.88646.554.114Jan 12, 2026

Evaluation results

MethodLinks
VST-7B-SFTData Scale=4.1M, Training Phases=1 (S)2026.0160.67244.474.368.359.755.844.965.2
SmoothOp-7BData Scale=50K, Training Phases=1 (R)2026.016071.141.472.270.361.755.142.865.4
VST-3B-SFTData Scale=4.1M, Training Phases=1 (S)2026.0157.969.345.471.862.4594638.770.2
SmoothOp-3BData Scale=50K, Training Phases=1 (R)2026.0153.970.142.869.938.360.445.438.765.5
InternVL-Spatial-8BData Scale=291K, Training Phases=1 (S)2026.0152.368.740.963.154.347.7—29.960.5
SpatialLadder-3BData Scale=26K, Training Phases=3 (S+R)2026.0145.7————————
Gemini-1.5 ProData Scale=-, Training Phases=-2026.0145.456.230.964.143.651.346.33634.6
ViLaSR-7BData Scale=81K, Training Phases=2 (S+R)2026.0145.463.534.460.630.948.945.230.449.2
SpaceR-7BData Scale=151K, Training Phases=2 (S+R)2026.0143.561.928.660.935.238.24631.445.6
LLaVA-Video-7BData Scale=-, Training Phases=-2026.0135.648.51447.824.243.542.43430.6
GPT-4oData Scale=-, Training Phases=-2026.013446.25.343.838.23741.331.528.5
Qwen2.5-VL-7BData Scale=-, Training Phases=-2026.0132.734.519.447.640.832.824.532.529.4
LLaVA-OneVision-7BData Scale=-, Training Phases=-2026.0132.447.720.247.412.342.535.229.424.4
SAT-7BData Scale=175K, Training Phases=1 (S)2026.01—————47.341.137.136.1