Vertex AI on Google Distributed Cloud (GDC) air-gapped features a growing list of foundation Generative AI models you can test, deploy, and implement for your air-gapped applications. Foundation models are fine-tuned for specific use cases and offered at different prices. This page summarizes the model families available in the Generative AI APIs on GDC and guides you on which models to choose by use case.
Embeddings models
Embeddings convert textual data written in a natural language into numerical
vectors. These vector representations are designed to capture the semantic
meaning and context of the words they represent. Text embedding models can
generate optimized embeddings for various task types, such as document
retrieval, questions and answers, classification, and fact verification. For
English text, use text-embedding-004. For multilingual text, use
text-multilingual-embedding-002.
The following table summarizes the models available in the Embeddings API. For more information on embeddings, see Text embeddings.
| Model | Description | Specifications |
|---|---|---|
| Text Embedding ( text-embedding-004) |
Returns embeddings for English text inputs. | Max token input: 2,048. Embedding dimensions: less than 768. |
| Text Embedding Multilingual ( text-multilingual-embedding-002) |
Returns embeddings for text inputs of over 100 languages. | Max token input: 2,048. Embedding dimensions: less than 768. |