Perplexity Launches 190 Million Document AI Benchmark for Agentic RAG

The newly released leaderboard evaluates 13 embedding models on their ability to perform large-scale web searches using agent-reformulated queries. Perplexity