Model role and official documentation
How one representation space connects different media →
The first EmbeddingGemma focused on text.
Google DeepMind · Released 2026-10-06
EmbeddingGemma 2
Create comparable representations of text, photos, audio and video for on-device media retrieval.
Original model explanation ↗No comparable leaderboard score is included yet. Explore the evaluations and examples below.
Evaluations: conditions and results
How this capability connects to works and uses
Official phone-retrieval example
Describe a scene and find its timestamp →
In Google’s phone demo, a query about a turtle eating returns several candidate moments, each with a timestamp.
Google DeepMind model evolution
- Gemini 3.7 Flash → 3.8 Flash
Flash advances coding, multi-step reasoning and agent work at the same speed and price positioning; Flash Cyber is a separate security-focused variant.
Original release ↗ - Gemini 4 Argon
The new generation is announced with access initially limited to invited testers; evaluation results do not establish public availability.
Original release ↗ - Nano Banana 2.1
Image generation and conversational editing advance, emphasizing local edits, subject consistency and text rendering.
Original release ↗ - EmbeddingGemma 2
The retrieval model extends text representations to images, audio and video, with public weights and phone demos.
Original release ↗