Tech Meridian ← LIVE FEED
PROMY MERIDIAN RU

NEWS · MODELS · #3062

Cohere's Embed 5 (Pro/Fast) and Rerank 5 available via OpenRouter

OpenRouter now exposes Cohere's retrieval models — Embed 5 (Pro and Fast) and Rerank 5 — through a single API with unified billing and no extra SDKs. The integration supports a recommended RAG pipeline (index with Embed 5 Pro, query with Embed 5 Fast, then rerank with Rerank 5), calibrated relevance scores, long (32k-token) rerank context, and example Python code using the OpenRouter SDK and NumPy for a demo vector store.

KEY POINTS

  1. OpenRouter now exposes Cohere's retrieval models — Embed 5 (Pro and Fast) and Rerank 5 — through a single API with unified billing and no extra SDKs.
  2. The integration supports a recommended RAG pipeline (index with Embed 5 Pro, query with Embed 5 Fast, then rerank with Rerank 5), calibrated relevance scores, long (32k-token) rerank context, and example Python code using the OpenRouter SDK and NumPy for a demo vector store.
  3. Makes high-quality retrieval models easier to use in production RAG pipelines by removing additional SDK/setup friction and enabling a Pro-index/Fast-query pattern and calibrated reranking.

WHY IT MATTERS

Makes high-quality retrieval models easier to use in production RAG pipelines by removing additional SDK/setup friction and enabling a Pro-index/Fast-query pattern and calibrated reranking.

SOURCES & TIMELINE

1
01
COHERE FIRST-PARTY
Cohere retrieval models on OpenRouter

Access Cohere’s state-of-the-art retrieval models through a single API, with unified billing and no additional SDKs. Embed 5 and Rerank 5 are now live on OpenRouter. Here’s how to turn them into a working RAG pipeline in a few dozen lines of Python code through a single API . There are Cohere’s most powerful retrieval models yet, delivering leading performance on enterprise documents and complex retrieval tasks. If y…

↗