Multivariate Anomaly Detection on TSB-AD multivariate 180 series (test) benchmark leaderboard · SOTA2 Research

Multivariate Anomaly Detection on TSB-AD multivariate 180 series (test)

43.7AUC-PRAxonAD

Evaluation Results

Method	Links
AxonAD2026.04		43.7	82.5	49.3	85.9
JuRe2026.04		40.4	80.2	44.4	83.2
Stream-VAE2026.04		39.9	80.8	45	83.8
OmniAnomaly2026.04		37.2	74.4	42.4	77.7
USAD2026.04		36.3	73.8	41.2	77.1
WVAE2026.04		35.4	74.7	41.3	77.8
CNN2026.04		34.7	77	35.2	80.7
VASP2026.04		33.9	76.2	40.1	80.9
SISVAE2026.04		32.3	75.9	37.2	78.6
M2N22026.04		31.9	74	32.3	77.9
OFA2026.04		30	63.9	36.7	69.4
AutoEncoder2026.04		29.4	66.9	29.5	69.1
VSVAE2026.04		29	70.9	34.2	73.4
GDN2026.04		27.2	73.8	33.2	80.2
TranAD2026.04		25.8	67.5	30.8	74.2
KMeansAD2026.04		25.2	69.1	29.6	73.2
TFTResidual2026.04		25	71	30.8	77.7
LSTMAD2026.04		24.8	59.7	24.5	62.6
PCA2026.04		24.2	67.6	27.7	71.2
Donut2026.04		21.3	62.7	26.2	69.3
IForest2026.04		21	70.4	25.3	75
COPOD2026.04		20.5	65.2	20.3	68.6
TimesNet2026.04		20.1	61.8	27.1	68.6
FITS2026.04		19.7	61.1	26.7	68.6
LOF2026.04		9.6	53.4	13.8	59.7
AnomalyTr.2026.04		6.8	50.6	11.5	53.8