Google released EmbeddingGemma 2, an open multimodal embedding model small enough to run on a smartphone, adding image, audio and video support.
Google released EmbeddingGemma 2, an open model with 740 million parameters that converts text, images, video, audio, and code into vectors. According to Google, it outperforms some competing models twice its size and runs on-device using about 191 MB of RAM.
Why it matters: The model is small enough to run on-device with low memory use, which could make multimodal embedding tasks practical on local hardware.
Written by AI from The Decoder's published text.
Background
From our own funding, model and news trackers, not from the source.
- Newest model on the API
- Gemini 4 Argon · 13 Oct 2026 · $2 in / $10 out per million tokens
- Last 30 days
- 156 news stories
More coverage of this story (2)
Google releases EmbeddingGemma 2, an open lightweight multimodal embedding model
Primary source
“Google released EmbeddingGemma 2, an open model with 740 million parameters that converts text, images, video, audio, and code into vectors.”
Reporting from a publisher outside the organisation being covered.
Read original story ↗