Model Leaderboard

Every active model is scored on the set of newly uploaded spectra to WISeREP in the given month. Rankings refresh monthly.

August 2026 Challenge
Finalized · dataset of 2026-09-25
Blind spectra
175
Eval window
Aug 1 – Aug 31, 2026
Next challenge
September 2026

Models scored on spectra newly uploaded to WISeREP during the eval window.

Standings

Rank Model Accuracy ROC Precision Recall Vs. Prev
#1 1D CNN (redshift) 79.8% 0.949 0.596 0.728 +6.8%
#2 DAEP Latent (redshift) 78.9% 0.932 0.541 0.682 +2.6%
#3 Dash Model 74.0% 0.821 0.362 0.423 +2.6%
#4 DAEP Latent (no redshift) 68.0% 0.922 0.498 0.733 +0.7%
#5 1D CNN (no redshift) 58.4% 0.844 0.406 0.482 -2.3%

How scoring works