Embed Multilingual Light 3
Embed Multilingual Light 3 is the compact multilingual option in the v3 embedding family. Cohere provides API access and private deployment guidance; the repository distributes a tokenizer.
Model behavior
The official card documents multilingual training and search examples. Queries and documents use their respective input types. The model is distinct from the larger Multilingual 3 variant, so keep the selected encoder recorded with its index.Embed Multilingual Light 3 tokenizer and API guide
How to use it well
Practical guidance. Consider Light for short multilingual records when storage use is a key constraint. Review the languages with the least data in your collection as well as the most common ones. A single average score can hide a weak language group. Compare relevance at the number of results the interface actually shows, and retain enough source text for users to assess each match.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | embed-multilingual-light-v3.0 |
| Context window | 512 tokens Text input token limit for this v3 variant. Cohere Embed specifications |
| Vector dimensions | 384Cohere Embed specifications |
| Language scope | More than 100 languagesEmbed Multilingual Light 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 Multilingual 3
Cohere · Embed
Multilingual text embedding model for retrieval across languages.
Launch date not confirmedAPI