OpenRouter· Updated
OpenRouter introduced a dedicated API for reranker models, starting with the Cohere suite. While standard vector search finds similar text, rerankers score those results for actual relevance to ensure the LLM receives the highest-quality context. This update allows developers to manage both retrieval optimization and model inference through a single provider.
rerank-4-pro, rerank-4-fast, and rerank-v3.5.Standard embedding-based search finds relevant chunks, but rerankers determine which ones are most relevant to a query. These models act as a high-accuracy filter, identifying the best information before it reaches the model. Integrating these tools removes the need for separate infrastructure to handle the precision layer of a production RAG stack.
You can now implement reranking via the POST /api/v1/rerank endpoint by passing a query and document chunks. The available models support 100+ languages with no pre-processing. rerank-4-pro offers a 32K context window, while rerank-4-fast is optimized for applications requiring the lowest possible latency.
New on OpenRouter: Reranker Models 🔍 Why add a reranker to your RAG pipeline? Embedding search finds relevant chunks, but rerankers tell you which ones are MOST relevant and result in significantly better answers. Now live via a single API endpoint, starting with @cohere! https://t.co/ZE0vrLOHHl
Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →
More like this
