Embed 5 Fast
Embed 5 Fast targets frequent queries and applications that need short response times. It uses the same embedding space as Pro, so the query encoder can change without rebuilding a compatible Pro index.
Model behavior
The model accepts text, images and combined content. It is available through Cohere and supported hosting platforms, including OpenRouter. The OpenRouter guide notes that its embedding responses use floats, which matters when comparing that route with a workflow built around compressed vectors.Embed 5 and Rerank 5 on OpenRouter
How to use it well
Practical guidance. Fast fits the query side of an interactive search application. Measure the time from a user's question to useful results, including the vector store and any reranking step. If retrieval quality is weak, inspect whether the right document was indexed before changing the encoder. Keep the output dimension consistent between stored documents and incoming queries. Sharing a family name with an older Embed release does not establish index compatibility.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | embed-v5.0-fast |
| Context window | 128,000 tokens Maximum input context reported for Embed 5 Fast. Cohere Embed specifications |
| Model scale | Approximately 1 billion parametersCohere Embed specifications |
| Vector dimensions | 256, 512, 768, 1024, 1536 or 2048Cohere Embed specifications |
Related models
- Rerank 5 Fast
Cohere · Rerank
Low-latency search reranker with calibrated relevance scoring and Context Focus.
Launch date not confirmedAPI - Embed 5 Pro
Cohere · Embed
Quality-focused embedding model for text, images and mixed documents, sharing a vector space with Embed 5 Fast.
Launch date not confirmedAPI - Embed Multilingual Light 3
Cohere · Embed
Compact multilingual text embedding model in the v3 family.
Launch date not confirmedAPI