Back

Gemini Embedding Model Launches with Support for Over 100 Languages

The Gemini Embedding text model is now available to developers, offering advanced capabilities for diverse applications through the Gemini API and Vertex AI.

Summary

The Gemini Embedding text model (gemini-embedding-001) has been officially launched for developers via the Gemini API and Vertex AI, supporting over 100 languages with a maximum input length of 2048 tokens. Priced at $0.15 per million input tokens, it leverages Matryoshka Representation Learning technology to excel in fields such as science, law, finance, and programming. Developers can try the model for free through Google AI Studio.

Terms & Concepts
  • Gemini API: An application programming interface that allows developers to access and integrate Gemini's machine learning models into their applications.
  • Vertex AI: A Google Cloud service that enables developers to build, deploy, and scale machine learning models.
  • Matryoshka Representation Learning: An advanced learning technique used in the Gemini Embedding model to enhance its performance across various domains.