What modalities does Google say EmbeddingGemma 2 supports?
Google describes a 740-million-parameter multimodal embedder that places text, code, images, video and audio in a shared embedding space, with on-device inference as a target. Its benchmark leadership claim is scoped to the compared models and sizes, not a universal superiority result.
Answered in
EmbeddingGemma 2 Wants Your Photos, Audio and Code in the Same Search SpaceGoogle’s multimodal embedding announcement points toward cross-media retrieval, where matching meaning matters more than generating an answer.
Read the full analysisOther questions this article answers
More news questions
- Why is a candidate vulnerability count an incomplete security measure?
- Do the reported 29,000 candidate vulnerabilities mean 29,000 confirmed bugs?
- Why is vulnerability validation difficult for maintainers?
- Why distinguish model availability from a later documentation update?
- What does Apertus 1.5 support?
- Why publish intermediate checkpoints?
- Why should a search product distinguish retrieval from answer generation?
- How can teams test cross-modal search?
Every answer on Crashtech is written by the editor of the article it comes from — never auto-summarised. Browse all answers or the News beat.