{"type":"post","id":"google-released-embeddinggemma-2-a-740m-open-weight-8a4f42","slug":"google-released-embeddinggemma-2-a-740m-open-weight-8a4f42","title":"Google released EmbeddingGemma 2 , a 740M open-weight multimodal embedding model for text, image, audio, and video.","description":"EmbeddingGemma 2 makes shared multimodal retrieval available as a local model rather than requiring separate cloud or on-device stages for image, speech, and text processing.","retrieval_nugget":"EmbeddingGemma 2 makes shared multimodal retrieval available as a local model rather than requiring separate cloud or on-device stages for image, speech, and text processing.","published_at":"2026-10-21","updated_at":"2026-10-21","record_date":"2026-10-21","date_kind":"published_at","topics":["google"],"source_urls":["https://developers.googleblog.com/google-ai-edge-with-embeddinggemma-2"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:google-released-embeddinggemma-2-a-740m-open-weight-8a4f42","status":"published","visuals":[],"editorial_provenance":{"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:44bfd15c626ef5dca42e74aeee334fd4b9c1210896f4de60437eb137b2668810","reviewed_at":"2026-10-21T05:01:44.354Z","source_evidence_count":1,"verified_claim_count":4,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":4,"implication_count":1,"watch_condition_count":1,"related_record_count":0},"analysis":{"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"The thesis is that a compact shared representation can move some media search from model orchestration into an on-device retrieval layer.","observed_facts":[{"text":"Google announced EmbeddingGemma 2 as an open-weight 740-million-parameter model mapping text, images, video frames, and audio into a shared vector space.","source_urls":["https://developers.googleblog.com/google-ai-edge-with-embeddinggemma-2"]},{"text":"Google reports roughly 567 MB active RAM for the full multimodal model and 191 MB for text-only weights on a Pixel 11 Pro.","source_urls":["https://developers.googleblog.com/google-ai-edge-with-embeddinggemma-2"]},{"text":"Google AI Edge Gallery demos local photo search and timestamp retrieval in video without transcription or generated captions.","source_urls":["https://developers.googleblog.com/google-ai-edge-with-embeddinggemma-2"]},{"text":"Google says ML Kit support is planned for the coming weeks.","source_urls":["https://developers.googleblog.com/google-ai-edge-with-embeddinggemma-2"]}],"mechanism":"The mechanism is a single embedding space plus local indexing and similarity ranking, rather than a chain of modality-specific models and intermediate text.","why_now":"Why now: Google is showing a route from model release to user-facing local search demos, while saying ML Kit support is planned in the coming weeks.","implications":["The first implication is that teams can prototype private, offline media retrieval around one embedding interface instead of maintaining separate captioning, transcription, and text-embedding stages."],"evidence_boundary":"The evidence boundary is limited to the vendor’s described model, demos, and roadmap; no comparative quality, latency, battery, or device-compatibility results are supplied here.","watch_conditions":["Watch whether the planned ML Kit support ships and how retrieval quality, latency, and memory use vary across devices."],"related_records":[],"new_branch_reason":"No prior New Runtime record was supplied for this release, so this starts a distinct record focused on local multimodal retrieval as application infrastructure."},"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/google-released-embeddinggemma-2-a-740m-open-weight-8a4f42/","markdown":"https://newruntime.com/posts/google-released-embeddinggemma-2-a-740m-open-weight-8a4f42.md","json":"https://newruntime.com/posts/google-released-embeddinggemma-2-a-740m-open-weight-8a4f42.json"},"next_reads":[{"type":"topic","path":"/topics/google/","url":"https://newruntime.com/topics/google/","title":"Google","reason":"Explore the google topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/google.md","json_url":"https://newruntime.com/topics/google.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/source-ledger.json","url":"https://newruntime.com/source-ledger.json","title":"Source Ledger","reason":"Inspect the public source evidence graph.","media_type":"application/json"}]}