Google's first natively multimodal embedding model. Maps text, images, video, audio, and PDFs into a unified vector space. Supports Matryoshka Representation Learning with configurable dimensions from 128 to 3072. Covers 100+ languages. GA released April 22, 2026.
Parameters
N/A (embedding model)
Context Window
8192 tokens
License
Proprietary
Release Date
2026-04-22
API Pricing
API pricing for this model is not yet available