Cohere Embed 5 and Rerank 5 are now on OpenRouter (November 4, 2026)

What you get

On November 4, 2026 Cohere put Embed 5 and Rerank 5 on OpenRouter. If you already use OpenRouter, you can add Cohere embeddings and reranking with your existing key.

• Embed 5 takes text, images, and mixed text and images such as PDF pages, with a 128k-token context window.
• Embed 5 Pro and Fast share one embedding space. Index your documents with Pro and embed queries with Fast.
• Rerank 5 has a 32,768-token context window for each query and document combined. Cohere says Rerank 5 Fast beats Rerank 4 Pro.
• Prices are on each model's OpenRouter page. Cohere's announcement doesn't list them.

Prerequisites

• An OpenRouter API key in OPENROUTER_API_KEY
• The OpenRouter Python SDK (pip install openrouter, Python 3.10 or later), or any HTTP client

Steps

1. Pick a model ID from the table and pass it as model.
2. For embeddings, call OpenRouter's Embeddings API. It accepts input, input_type, dimensions and encoding_format (float or base64). Vectors come back in data[].embedding.
3. For reranking, call OpenRouter's Rerank API. It accepts query, documents and top_n. Each item in results[] has , and .