As of: 2026-11-22 13:32 CET. Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/ Canonical page: https://postcutoff.com/m/embeddinggemma-2/ # EmbeddingGemma 2 Google DeepMind, Gemma family. Status: current. Type: embedding model. Released: 2026-10-06. ## How to call it | Provider | Model id | Endpoint | Docs | |---|---|---|---| | Hugging Face | `google/embeddinggemma-2` | | https://huggingface.co/google/embeddinggemma-2 | | Google docs | | | https://ai.google.dev/gemma/docs/embeddinggemma | ## Pricing - Price not published. - Last checked: 2026-10-07 ## Specs - Input: text, image, video, audio - Output: text - Context window: 8,192 tokens - Max output: not published - Open weights: yes - Licence: apache-2.0 ## Where it stands - No newer embedding model in the registry yet. - Before it from Google: Gemini Embedding 2 (2026-04-22) https://postcutoff.com/m/gemini-embedding-2/ - Weights are public under a permissive licence (Apache-2.0). ## What stands out - Four modalities in one small open embedding space: Maps text (including code), images, video and audio into one 768-dimensional space with 740M parameters in total: a 270M text model (130M backbone + 140M embedder) plus optional 170M vision and 300M audio encoders that can be loaded separately. Source: https://huggingface.co/google/embeddinggemma-2 - Matryoshka truncation: Embeddings can be cut to 512, 256 or 128 dimensions (up to 6x less storage); quality loss is small down to 256d. Source: https://huggingface.co/google/embeddinggemma-2 - Big code-retrieval gain over v1: MTEB code v1 78.68 vs 68.76 for EmbeddingGemma 1; MTEB multilingual v2 61.36 vs 61.15; MMEB v2 overall 59.01 (Google's numbers). Source: https://huggingface.co/google/embeddinggemma-2 ## Notes Announced 2026-10-06 on the Google blog. Apache 2.0 (Gemma 4 license), not gated on Hugging Face. Runs on phones and laptops; uses task-instruction prefixes (e.g. 'task: search result | query: ...') for text inputs. sentence-transformers id: google/embeddinggemma-2. No hosted Gemini API endpoint is listed; for hosted multimodal embeddings Google offers gemini-embedding-2. Small open on-device embedding model for multimodal search, RAG, classification and clustering. ```python from sentence_transformers import SentenceTransformer model = SentenceTransformer("google/embeddinggemma-2") q = model.encode("What causes the northern lights?", prompt_name="SearchQuery") d = model.encode("The northern lights are caused by charged particles from the sun.", prompt_name="Document") print(model.similarity(q, d)) ``` Sources: [Hugging Face model card](https://huggingface.co/google/embeddinggemma-2), [launch blog](https://blog.google/innovation-and-ai/technology/developers-tools/embeddinggemma-2/), [docs](https://ai.google.dev/gemma/docs/embeddinggemma). Model page: https://postcutoff.com/m/embeddinggemma-2/