BeDraft's Darwin-27B-ZTC-v5 tops S1MB benchmark, beating Perplexity and Meta

BeDraft's Darwin-27B-ZTC-v5 ranked first among 152 models on the open S1MB benchmark with a Borda score of 91.16. It beat Perplexity (88.66) by 2.5 points and TypeSafe AI's official Jev 1.13.0 model (14th place) by 9.72 points. Two BeDraft models made the top five.

• Darwin-27B-ZTC-v5 scored 91.16 on Borda versus Perplexity's 88.66
• It beat Jev 1.13.0 by 9.72 points overall
• The Score category gap reached 18.34 points
• It won 97 of 137 head-to-head tasks against Jev