Perplexity's 9B and 0.6B embedding models beat nemotron-embed-8b on Q2D-Web
Original titleOn Q2D-Web, about 70,000 production-derived search queries over 190M web documents, the 9B and 0.6B models reach 74.8% and 73.6% Recall@1...
AISummary
On the Q2D-Web benchmark of about 70,000 production-derived search queries over 190M web documents, Perplexity's 9B and 0.6B models reach 74.8% and 73.6% Recall@1000. Both outperform nemotron-embed-8b, which scores 69.3%.
Source: Perplexity · x.comPublished · added here