Model-based Rankers

Model Ranker transforms Zilliz Cloud search by integrating advanced language models that understand semantic relationships between queries and documents. Instead of relying solely on vector similarity, it evaluates content meaning and context to deliver more intelligent, relevant results.

Cohere Ranker [READ MORE]

The Cohere Ranker leverages Cohere's rerank models to improve result ordering by applying semantic reranking to retrieved candidates.

Voyage AI Ranker [READ MORE]

The Voyage AI Ranker leverages Voyage AI's and search applications.

Hugging Face Ranker [READ MORE]

Vector search orders results by vector distance, but the initial order may not reflect how well each candidate's text answers the query. With a Hugging Face model provider integration, Hugging Face Ranker uses scores from the Hugging Face sentence-similarity task to reorder the candidates returned by vector search.