EmbeddingGemma 2
Free, tiny Google search model for text, images, audio and video; scores are Google's own, so wait for outside results.
Should you use it?
Pick it if
Developers building search or retrieval that must run on a phone or laptop, or across mixed media.
Skip it if
You need text-only search with proven results; version 1 scores almost the same on multilingual text s1.
The short version
- What this is
- An embedding model from Google: it turns text, code, images, video and audio into vectors so apps can search across them. Free weights, small enough to run on a phone.
- Why it matters
- Search quality depends on embeddings; this one is free and runs offline, but its quality claims are not yet checked by anyone outside Google.
- Free under Apache 2.0, 740M parameters, runs on device s1 s2.
- Google reports a big gain on code search and almost none on multilingual text s1.
- No independent benchmark yet, so no score yet.
Good at
Our take
EmbeddingGemma 2 is a free model that lets an app search text, pictures, video and sound together, and it is small enough for a phone s1 s2. That is useful if your data can't leave the device.
The authors report a large jump on code search, 78.68 against 68.76 for version 1 on MTEB code, but almost no change on multilingual text s1. Those are Google's own numbers. If you only search text, version 1 looks nearly as good on that table. We'll score it once an outside benchmark lists it.
Not scored yet
This is a First look: the call above comes from the maker's own pages, so there is no score yet. Once MTEB or another independent embedding benchmark publishes its own results for it.
Research
New to this? What the scores and labels mean
- 0 to 100
- Our score for one job, like writing or coding, compared with everything else out today. 100 is the best available today. For AI models we build it from public data, such as Arena's blind votes and Epoch AI's independent tests; apps and tools are scored on our own checklist. The same tool can score differently for different jobs, and where two likely ranges overlap, we call it a tie.
- Our verdict
- Use it: a top choice for most people. Worth trying: good, and worth a go. Depends: good for some people or one kind of job. Wait: too early to tell, or not out for everyone yet. Skip: something else does the job better.
- First look
- Just released. We say what it is and whether to bother, and give it a score once public results exist.
- No score yet
- None of the public sources we use covers it yet, or the only evidence is from a source whose numbers we can't use, so we describe it in words instead.
- Best buy
- Nothing cheaper scores as well for that job.
- Best free
- The best score for that job that costs nothing, through a free plan or by running on your own computer.
- Pay as you go
- No monthly plan: you pay for each use, usually through a developer account.
- Checked
- The day we last confirmed the score and the price against the sources. Prices change often, so check the date.
- AI model
- The engine underneath, like GPT or Claude. You use it through an app or, if you code, directly.
- Open source
- Free to download and run on your own computer, so nothing leaves your machine.
Alternatives
Better or cheaper for research, from our scores, then the ones we'd also look at. The same kind of product comes first.
- Gemini 4 ArgonTop scoredText model95$4.99 a monthScores 95 in research, first in our ranking; $4.99 a month. Side by side
- Cohere Embed 5 ProEmbedding modelFirst look: WaitNo public priceHosted embedding model for text, images and PDFs, also waiting on independent scores. Side by side
- OllamaPlatform80CodingFreeFree way to run it locally on your own computer. Side by side
Evidence
Everything behind the call and the scores, folded so the page stays short. Spotted a wrong price, score or link? Report a mistake.
How we scored this
Reviews are drafted with help from AI models, including Anthropic's Claude. Model scores come only from outside leaderboards and independent evaluations, by a written rule.
No one paid for this review or gave us free access, and we earn nothing if you sign up. How we score
Sources 2
- s1Hugging Face: google/embeddinggemma-2 model card
- s2SiliconANGLE: Google expands EmbeddingGemma beyond text to images, audio and video
Something here out of date or wrong? Report a mistake and we will check it against the source.
History 1 change
Every change to a rating, the price or our call, with the reason. Scores are against what exists on the day, so they drift down as better things ship. Older entries come from before we moved to scores.
- First look. Only Google's own benchmarks so far, so not rated.
Use it well 4 guides on this site
Pages on this site that help with the jobs above.