// AI Platform Drift
Cloudflare adds EmbeddingGemma 2 multimodal embedding model to Workers AI
Cloudflare developer changelog · published 2026-11-14 18:06 UTC
WHAT CHANGEDCloudflare Workers AI now offers @cf/google/embeddinggemma-2, a multimodal embedding model that maps text, images, video, and audio into a single 768-dimensional vector space. The model supports flexible output dimensions (128, 256, or 512), an 8,192 token context window, task prefixes for tuning, and over 100 languages. Pricing is $0.0017 per million input tokens.
WHY IT MATTERSTeams can now build cross-modal search and retrieval systems on Workers AI without maintaining separate embedding models for each content type. The shared vector space enables querying mixed collections of documents, products, and media with a single text query.
WHAT TO DOReview the EmbeddingGemma 2 model page at https://developers.cloudflare.com/workers-ai/ to evaluate whether multimodal embeddings fit your use case, particularly if you currently handle text and media separately or need multilingual code retrieval.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## Workers AI - EmbeddingGemma 2 now available on Workers AI +https://developers.cloudflare.com/changelog/post/2026-11-12-embeddinggemma-2-workers-ai/ +Thu, 12 Nov 2026 00:00:00 GMT + +@cf/google/embeddinggemma-2 is now available on Workers AI. EmbeddingGemma 2 is an open multimodal embedding model from Google DeepMind. It maps text (including code), images, video, and audio into a single 768-dimensional vector space, so you can search a mixed collection of products, documents, and media with one text query. +Key capabilities: + +Multimodal: Embed text, images, video, audio, or any combination of them in one input. Every embedding shares the same vector space. +Multilingual and code: Understands more than 100 languages and improves code retrieval over the original EmbeddingGemma. +Flexible dimensions: Set dimensions to 128, 256, or 512 to return shorter vectors and reduce storage, with a small cost to quality. +Long context: An 8,192 token context window, shared across all content in an input. +Task prefixes: Add short instructions such as task: search result | query: ... to tune embeddings for search, classification, clustering, or similarity. + +Send text with input. Send images, video, and audio as base64 data URIs in messages: +const response = await env.AI.run("@cf/google/embeddinggemma-2", { +messages: [ +[ +{ +role: "user", +content: [ +{ type: "text", text: "Waterproof running shoes." }, +{ type: "image_url", image_url: { url: imageDataUri } }, +], +}, +], +[ +{ +role: "user", +content: +"task: search result | query: waterproof shoes for trail running", +}, +], +], +});const response = await env.AI.run("@cf/google/embeddinggemma-2", { +messages: [ +[ +{ +role: "user", +content: [ +{ type: "text", text: "Waterproof running shoes." }, +{ type: "image_url", image_url: { url: imageDataUri } }, +], +}, +], +[ +{ +role: "user", +content: +"task: search result | query: waterproof shoes for trail running", +}, +], +], +}); +Pricing is $0.0017 per million input tokens. +For more information, refer to the EmbeddingGemma 2 model page and pricing.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | New embedding model with distinct capabilities, pricing, and API surface added to Workers AI platform. |
| signals | new model added to platform · pricing specified: $0.0017 per million input tokens · new API capability: multimodal embeddings (text, image, video, audio) · new parameter options: flexible dimensions (128, 256, 512) · new context window: 8,192 tokens · new feature: task prefixes for tuning |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |