Embed English 3
Embed English 3 is an English embedding model accessible through Cohere's API and private deployment routes. Its public repository contains the tokenizer, not downloadable model weights.
Model behavior
The repository demonstrates distinct input types for document and query encoding. That distinction is part of the retrieval setup, rather than a choice of how the final answer should be written.Embed English 3 tokenizer and API guide
How to use it well
Practical guidance. Use this model as a baseline for an existing English archive with short passages. Compare exact-name queries, paraphrases and documents with similar titles. When a useful answer spans several passages, retain links between those passages and the source document. Choose a newer long-context model only after checking whether passage boundaries, rather than the encoder, cause the missed matches.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | embed-english-v3.0 |
| Context window | 512 tokens Text input token limit for this v3 variant. Cohere Embed specifications |
| Vector dimensions | 1024Cohere Embed specifications |
| Language scope | EnglishEmbed English 3 tokenizer and API guide |
Related models
- Embed 5 Fast
Cohere · Embed
Low-latency embedding model that shares its vector space with Embed 5 Pro.
Launch date not confirmedAPI - Embed English Light 3
Cohere · Embed
Compact English embedding model in Cohere's v3 family.
Launch date not confirmedAPI