embeddinggemma-300m

google Gemma Terms of Use out 15 months ago

303M parameters, text in, embeddings out, 2K tokens of context. 2 providers serve it from $0.002 per million input tokens.

Verified 2026-10-25 Rank #6 in Embeddings · score 109.3 method v1 Sources 3 Downloads 30d 2.4M · #58 of 619 tracked

Changed this week: 4 new quantized builds on the Hub (321 in all).2026-10-24

Can I run it?0.2GB at INT4Estimatefits a 16 GB cardmethod

What does it cost?$0.002/1M incheapest of 1DeepInfra2026-10-24

Can I use it?Gemma Terms of UseGoogle's terms + prohibited-use policy; must pass terms downstreamlicence file2026-10-24

Does it work?69.67on MTEB#6 of 64 measuredleaderboard2026-10-24

Do people use it?2.4Mdownloads, 30 days#58 of 619 we track · top 9%+25 likes, downloads −19% this week2,056 likesHugging Face2026-10-24

Worth reading and watching

Picked by rule from 18 tracked pieces: the lab's own post, the most shared, a field report, the latest.

Most discussed · 11 HN pts

Google EmbeddingGemma 2

twitter.com 2026-10-06 11 HN pts · 1 comments about the embeddinggemma family family

Latest · 19 days ago

EmbeddingGemma 2: An open model for natively multimodal embeddings [video]

youtube.com 2026-10-06 2 HN pts about the embeddinggemma family family

18 best of 18 tracked

The 2 picks above, plus these 16. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

ArticleEmbeddingGemma 2: An open, lightweight multimodal embedding modelblog.google · 2026-10-06 · 378 HN pts · official
ArticleEmbeddingGemma 2developers.googleblog.com · 2026-10-06 · 9 HN pts · official
ArticleEmbeddingGemma - Google AI for Developersai.google.dev · undated · official
ArticleEmbeddingGemma: The Best-in-Class Open Model for On-Device Embeddingdevelopers.googleblog.com · 2025-09-04 · 44 HN pts · official
PaperEmbeddingGemma: Powerful and Lightweight Text ...arxiv.org · undated
ArticleEmbeddingGemma: Architecture and Recipedevelopers.googleblog.com · 2025-09-29 · 1 HN pts · official
ArticleWelcome EmbeddingGemma, Google's new efficient embedding modelhuggingface.co · 2025-09-04 · 3 HN pts
ArticleEmbeddingGemma Model Cardai.google.dev · 2025-09-07 · 2 HN pts · official
CodeClient side Vector Search using EmbeddingGemma – WebGPU acceleratedgithub.com · 2026-03-24
CodeShow HN: Sift – Score your feed locally with EmbeddingGemmagithub.com · 2026-02-24
CodeShow HN: RAXE Open Source – LLM Prompt Threat Detection (EmbeddingGemma L2)github.com · 2025-12-20
Article[Literature Review] EmbeddingGemma: Powerful and Lightweight Text Representationsthemoonlight.io · undated
ArticleBuilding an Embedding API with Rust, Arm, and EmbeddingGemma on AWS Lambdasobolev.substack.com · 2026-02-13
VideoHow to Fine-Tune Google's EmbeddingGemma – LLMyoutube.com · 2025-09-06
ArticleEmbeddingGemma: Powerful and Lightweight Text Representations — Pith Reviewpith.science · 2025-09-24
ArticleEmbeddingGemma: Powerful and Lightweight Text Representations [Quick Review]liner.com · 2025-09-24

Show all 16

Where to run it

Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.

#	Cheapest 2	In /1M	Out /1M	Context	Quant	Seen
1	DeepInfra	$0.002	—	2K	—	Oct 24
2	HF Inference	—	—	—	—	Oct 25

All 2 providers →

Where it's used

Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.

Served by 2 API providers; heavily run locally (321 quantized builds); a big base for builders (305 fine-tunes, adapters and merges).

As a service2DeepInfra, HF Inference
On own hardware321quantized builds
Built on305293 fine-tunes · 12 adapters · 0 merges
In apps100+Hugging Face Spaces

Top Spaces running it: embedl/hfviewer ♥ 181 · google/mood-palette ♥ 138 · google/embeddinggemma-tuning-lab ♥ 20 · jairwaal/image ♥ 16

Will it fit?

Weights only, an estimate: leave room for the KV cache. Method

Fits a 16 GB card at INT4

About 0.2 GB at INT4, 0.4 GB at FP8, 0.7 GB at BF16: weights only, so leave room for the KV cache.

Estimate params × bytes per param × 1.2 · how we estimate

How it scores

Third-party measurements only; author-reported numbers stay in Reference, marked as such.

#6 of 64 on MTEB so far

69.67. Each tick is a model on this site with a measured MTEB row; author-reported numbers stay in Reference. Mean(task) x100 over 41 tasks. 41 tasks, split=test

Source · leaderboard · 2026-10-24

The story so far

Weights, every dated article, and repo commits on one rail.

Most coverage arrived in the last 30 days

Teal is the weights (2025-07-17), amber triangles are 8 dated articles (each links to it), grey ticks are repo commits.

What people built on it

626 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

626 derivatives on the Hub, most of them quantized

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

GGUF ggml-org/embeddinggemma-300M-GGUF for llama.cpp, Ollama, LM Studio 8.6M downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

Benchmarks

45 rows

Benchmarks 45 rows

measured third-party leaderboard row (3)author-reported number from the model card or eval files (42)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.

Benchmark	Value	Setting / notes	Reported by	Date
measured MTEB · Embed text for search	55.69 nDCG@10 x100 (mean of 10 tasks)	10 tasks, split=test	leaderboard	2026-10-24
measured MTEB · Embed text for search	69.67 mean(task) x100 over 41 tasks	41 tasks, split=test	leaderboard	2026-10-24
measured MTEB Reranking · Rerank search results	47.43 MAP x100 (mean of 2 tasks)	2 tasks, split=test	leaderboard	2026-10-24
author-reported MTEB · Embed text for search	61.15	768d	model author	2026-10-07
author-reported MTEB · Embed text for search	54.31	768d	model author	2026-10-07
author-reported MTEB · Embed text for search	60.71	512d	model author	2026-10-07
author-reported MTEB · Embed text for search	53.89	512d	model author	2026-10-07
author-reported MTEB · Embed text for search	59.68	256d	model author	2026-10-07
author-reported MTEB · Embed text for search	53.01	256d	model author	2026-10-07
author-reported MTEB · Embed text for search	58.23	128d	model author	2026-10-07
author-reported MTEB · Embed text for search	51.77	128d	model author	2026-10-07
author-reported MTEB · Embed text for search	69.67	768d	model author	2026-10-07
author-reported MTEB · Embed text for search	65.11	768d	model author	2026-10-07
author-reported MTEB · Embed text for search	69.18	512d	model author	2026-10-07
author-reported MTEB · Embed text for search	64.59	512d	model author	2026-10-07
author-reported MTEB · Embed text for search	68.37	256d	model author	2026-10-07
author-reported MTEB · Embed text for search	64.02	256d	model author	2026-10-07
author-reported MTEB · Embed text for search	66.66	128d	model author	2026-10-07
author-reported MTEB · Embed text for search	62.7	128d	model author	2026-10-07
author-reported MTEB · Embed text for search	68.76	768d	model author	2026-10-07
author-reported MTEB · Embed text for search	68.76	768d	model author	2026-10-07
author-reported MTEB · Embed text for search	68.48	512d	model author	2026-10-07
author-reported MTEB · Embed text for search	68.48	512d	model author	2026-10-07
author-reported MTEB · Embed text for search	66.74	256d	model author	2026-10-07
author-reported MTEB · Embed text for search	66.74	256d	model author	2026-10-07

Show all 45 rows (20 more)

Benchmark	Value	Setting / notes	Reported by	Date
author-reported MTEB · Embed text for search	62.96	128d	model author	2026-10-07
author-reported MTEB · Embed text for search	62.96	128d	model author	2026-10-07
author-reported MTEB · Embed text for search	60.62	Q4_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	53.61	Q4_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	60.93	Q8_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	53.95	Q8_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	60.69	Mixed Precision* (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	53.82	Mixed Precision* (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	69.31	Q4_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	64.65	Q4_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	69.49	Q8_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	64.84	Q8_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	69.32	Mixed Precision* (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	64.82	Mixed Precision* (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	67.99	Q4_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	67.99	Q4_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	68.7	Q8_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	68.7	Q8_0 (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	68.03	Mixed Precision* (768d)	model author	2026-10-07
author-reported MTEB · Embed text for search	68.03	Mixed Precision* (768d)	model author	2026-10-07

What the card says

2 cited sentences

Summary generated, cited

The model produces numerical vector representations of text with a maximum input context length of 2048 tokens.^[1] It is released under the Gemma Terms of Use and supports an output embedding dimension of 768.^[2]^[3]

“Maximum input context length of 2048 tokens”
license
“Output embedding dimension size of 768”

Generated 2026-10-07 14:59 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.

Use-case evidence

3 use cases

What it is good at 3 use cases

Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.

Facts

21 fields

Listed in Embeddings

Facts

embeddinggemma-300m has 303M parameters. (safetensors metadata, observed 2026-10-24)
Context length is 2,048 tokens (2K). (config.json max_position_embeddings, observed 2026-10-24)
License: Gemma Terms of Use (friction: attribution). Google's terms + prohibited-use policy; must pass terms downstream. (license file, observed 2026-10-24)
Created 2025-07-17, last modified 2025-09-25. (HF API, observed 2026-10-24)
Access is gated on Hugging Face (manual). (HF API, observed 2026-10-24)

Pipeline tag	sentence-similarity
Library	sentence-transformers
Architectures	Gemma3TextModel
Model type	gemma3_text
Dtype	float32
Params by dtype	F32: 302,863,104
Size band	S
MoE	no
Context length	2,048 (config.json max_position_embeddings)
Gated	manual
Papers	arXiv:2509.20354
License	Gemma Terms of Use (gemma)
Hub snapshot	2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · Gemma3TextModel · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Gemma3TextModel · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed

vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for embeddinggemma-300m: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)	0.7 GB · Edge (<4 GB)
FP8 (1 byte/param)	0.4 GB · Edge (<4 GB)
INT4 (0.5 bytes/param)	0.2 GB · Edge (<4 GB)

Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.

All variants

40 variants

Lineage

Original model Family: google/embeddinggemma 626 derivatives on the Hub

quantized: 321finetune: 293adapter: 12

Top variants (40 shown, ranked by downloads)

Variant	Relation	Quant	Params	Band	License	Author	Likes	Downloads
ggml-org/embeddinggemma-300M-GGUF	quantized	GGUF	—	—	—	ggml-org	52	8.6M
onnx-community/embeddinggemma-300m-ONNX	quantized	—	—	—	gemma	onnx-community	74	1.6M
farbodtavakkoli/OTel-Embedding-300M	finetune	—	303M	S	apache-2.0	farbodtavakkoli	0	1.2M
sentence-transformers/embeddinggemma-300m-medical	finetune	—	303M	S	apache-2.0	sentence-transformers	52	167.4k
unsloth/embeddinggemma-300m-GGUF	quantized	GGUF	—	—	gemma	unsloth	90	141k
ente-ai/embeddinggemma-300m-gguf	quantized	gguf	—	—	gemma	ente-ai	0	65k
trt19/gemma-300m-legal-ft1	finetune	—	303M	S	—	trt19	0	45.2k
emillykkejensen/EmbeddingGemma-Scandi-300m	finetune	—	303M	S	apache-2.0	emillykkejensen	2	44.4k
tardellirs/colibri-embed-ptbr	finetune	—	151M	S	gemma	tardellirs	105	36.9k
admiralakber/embeddinggemma-300m-Q4_K_M-GGUF	quantized	GGUF	—	—	gemma	admiralakber	1	16.5k
second-state/embeddinggemma-300m-GGUF	quantized	GGUF	—	—	gemma	second-state	0	12.1k
AROY76/Embedding-gemma-300M-skills	finetune	—	303M	S	—	AROY76	1	10.1k
gaianet/embeddinggemma-300m-GGUF	quantized	GGUF	—	—	gemma	gaianet	0	9.4k
tardellirs/embeddinggemma-pt-br	finetune	—	152M	S	gemma	tardellirs	1	8.1k
mradermacher/embeddinggemma-300m-GGUF	quantized	GGUF	—	—	gemma	mradermacher	1	8.1k
taide/embeddinggemma-GTAIDE-300m-2605	finetune	—	303M	S	other	taide	13	6.3k
ktcapraz/embeddinggemma-german_qna-checkpoint-500	finetune	—	303M	S	—	ktcapraz	1	6.5k
NorskHelsenett/eti-embeddinggemma-v1	finetune	—	303M	S	gemma	NorskHelsenett	1	6.4k
FastFlowLM/Embedding-Gemma-300M-NPU2	finetune	—	—	—	gemma	FastFlowLM	0	4.6k
cstr/embeddinggemma-300m-GGUF	quantized	GGUF	—	—	gemma	cstr	0	4.6k
Omartificial-Intelligence-Space/AraGemma-Embedding-300m	finetune	—	303M	S	apache-2.0	Omartificial-Intelligence-Space	15	3.1k
qhoxie/embeddinggemma-model2vec-256d	finetune	—	65.5M	S	mit	qhoxie	5	3.1k
cduk/embeddinggemma-300m-GGUF-with-dense-modules	quantized	GGUF	—	—	gemma	cduk	0	2.2k
Baragi-AI/Munche-768	adapter	—	—	—	gemma	Baragi-AI	20	823
deepmodal/embeddinggemma-300m-ko	finetune	—	303M	S	apache-2.0	deepmodal	0	1.6k
s3dev-ai/embeddinggemma-300m-gguf	quantized	gguf	—	—	—	s3dev-ai	1	1.5k
wilfredomartel/GemmaEmbedding-TripLegalCL-300k	finetune	—	303M	S	apache-2.0	wilfredomartel	0	1.5k
Bhuvneesh/embeddinggemma-300m-Q8_0-GGUF	quantized	Q8_0	—	—	gemma	Bhuvneesh	0	1.4k
sabafallah/embeddinggemma-300m-sentence-transformers-gguf	quantized	gguf	—	—	gemma	sabafallah	0	1.4k
AROY76/embedding-gemma-300m-job-titles	finetune	—	303M	S	—	AROY76	0	1.3k
ByronLeeee/EmbeddingGemma-300M-LawVault	finetune	—	303M	S	apache-2.0	ByronLeeee	1	992
metricspace/embeddinggemma-300m-q4k	quantized	GGUF	—	—	gemma	metricspace	0	975
mikeee/m2v-gemma-embedding-300m	finetune	—	65.7M	S	mit	mikeee	1	915
headwAI/embeddinggemma-300m	finetune	—	303M	S	gemma	headwAI	0	929
sensiarion/embeddinggemma-300m-code-8L-distill-int8	quantized	int8	—	—	gemma	sensiarion	0	892
Saidakmal/uz_embeddinggemma-300m	finetune	—	303M	S	—	Saidakmal	5	622
Surpem/Supertron-embedding-300M	finetune	—	303M	S	apache-2.0	Surpem	3	719
vectorranger/embeddinggemma-300m-medical-300k	finetune	—	303M	S	apache-2.0	vectorranger	3	686
SandLogicTechnologies/EmbeddingGemma-300m-GGUF	quantized	GGUF	—	—	gemma	SandLogicTechnologies	1	779
hreyulog/embedinggemma_arkts	finetune	—	303M	S	—	hreyulog	0	829

Repository changelog

4 commits

changed in the last 7 daysOther: 2026-10-24all changes

Repository changelog 4 commits

Date	Commit	SHA
2025-09-25	Update README.md	57c266a
2025-09-25	Update README.md	671e8c1
2025-09-10	Add notebook for "Use this Model" with Google Colab option (#9)	c5cfa06
2025-09-04	Super-squash branch 'main' using huggingface_hub	64614b0

All providers

2 providers · 2 endpoints

Where to run 2 providers · 2 endpoints

Hosted endpoints observed for embeddinggemma-300m, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.0020/M.

Provider	Model ref	Context	Quant	$/M in	$/M out	tok/s	TTFT	Tools	Status	Observed
DeepInfra deepinfra	google/embeddinggemma-300m	2K	—	$0.0020	—	—	—	—	live	2026-10-24
HF Inference hf	google/embeddinggemma-300m	—	—	—	—	—	—	—	live	2026-10-25

Articles and coverage

18 pieces

Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.

Official6

EmbeddingGemma 2: An open, lightweight multimodal embedding model

blog.google2026-10-06officialauthorityHN front pageabout the embeddinggemma familyHN 378 pts · 38 comments678 words

Google launches EmbeddingGemma 2, an open multimodal model expanding beyond text to unify code, images, video, and audio. Built on the Gemma 4 architecture with 740 million parameters and an Apache 2.0 license, it achieves a 9.92-point improvement on code performance. The model features an 8K token context window and supports dynamic vector truncation for storage efficiency.written, reviewed, cited

Developers can dynamically truncate output vectors from 768 dimensions down to 512, 256, or 128 dimensions.
The model processes up to 5.5 minutes of audio, 29 images, or 58 video frames within its 8K token context window.

EmbeddingGemma 2

developers.googleblog.com2026-10-06officialauthorityabout the embeddinggemma familyHN 9 pts · 0 comments1,271 words

EmbeddingGemma 2 is a compact open model under the Apache 2.0 license that maps text, code, images, video, and audio into a unified 768-dimensional space. Based on Gemma 4, this sub-1B model uses Matryoshka Representation Learning (MRL) to enable dynamic truncation down to 128 dimensions, achieving a 6x reduction in vector database storage requirements while retaining much of the original quality.written, reviewed, cited

Video is sampled at 1 frame per second by default, and audio should be 16 kHz mono.
At 128 dimensions, image, video, and speech retrieval quality drops to around 75%.
EmbeddingGemma 2 scores 14% higher than EmbeddingGemma 1 on MTEB (Code).

EmbeddingGemma - Google AI for Developers

ai.google.devofficialauthorityabout the embeddinggemma family370 words

EmbeddingGemma 2 is a 740M parameter, open-source model licensed under the Apache 2.0 License that maps text, images, audio, and video into a unified 768-dimensional vector space. The model supports up to 8,192 tokens and utilizes Matryoshka Representation Learning to deliver up to a 6x reduction in vector database storage costs. It enables demanding cross-modal workflows like Retrieval-Augmented Generation locally on consumer GPUs or CPUs.written, reviewed, cited

EmbeddingGemma: The Best-in-Class Open Model for On-Device Embedding

developers.googleblog.com2025-09-04officialauthorityabout the embeddinggemma familyHN 44 pts · 4 comments912 words

EmbeddingGemma: Architecture and Recipe

developers.googleblog.com2025-09-29officialauthorityabout the embeddinggemma familyHN 1 pts · 0 comments1,274 words

EmbeddingGemma Model Card

ai.google.dev2025-09-07officialauthorityabout the embeddinggemma familyHN 2 pts · 0 comments1,648 words

Analysis3

EmbeddingGemma 2: An open model for natively multimodal embeddings [video]

youtube.com2026-10-06about the embeddinggemma familyHN 2 pts · 0 comments

EmbeddingGemma: Powerful and Lightweight Text ...

arxiv.orgauthorityabout the embeddinggemma family376 words

Welcome EmbeddingGemma, Google's new efficient embedding model

huggingface.co2025-09-04authorityabout the embeddinggemma familyHN 3 pts · 1 comments5,896 words

Community1

Google EmbeddingGemma 2

twitter.com2026-10-06about the embeddinggemma familyHN 11 pts · 1 comments

All coverage tracked for this model (18) — append-only, nothing is ever removed

Title	Source	Published	First seen	Signals
EmbeddingGemma - Google AI for Developers	ai.google.dev official	—	2026-10-07	exa
EmbeddingGemma: Powerful and Lightweight Text ...	arxiv.org	—	2026-10-07	exa
[Literature Review] EmbeddingGemma: Powerful and Lightweight Text Representations	themoonlight.io	—	2026-10-07	exa
Google EmbeddingGemma 2	twitter.com	2026-10-06	2026-10-07	hn
EmbeddingGemma 2: An open model for natively multimodal embeddings [video]	youtube.com	2026-10-06	2026-10-07	hn
EmbeddingGemma 2	developers.googleblog.com official	2026-10-06	2026-10-07	hn
EmbeddingGemma 2: An open, lightweight multimodal embedding model	blog.google official	2026-10-06	2026-10-07	hn
Client side Vector Search using EmbeddingGemma – WebGPU accelerated	github.com	2026-03-24	2026-10-07	hn
Show HN: Sift – Score your feed locally with EmbeddingGemma	github.com	2026-02-24	2026-10-07	hn
Building an Embedding API with Rust, Arm, and EmbeddingGemma on AWS Lambda	sobolev.substack.com	2026-02-13	2026-10-07	hn
Show HN: RAXE Open Source – LLM Prompt Threat Detection (EmbeddingGemma L2)	github.com	2025-12-20	2026-10-07	hn
EmbeddingGemma: Architecture and Recipe	developers.googleblog.com official	2025-09-29	2026-10-07	hn, exa
EmbeddingGemma: Powerful and Lightweight Text Representations — Pith Review	pith.science	2025-09-24	2026-10-07	exa
EmbeddingGemma: Powerful and Lightweight Text Representations [Quick Review]	liner.com	2025-09-24	2026-10-07	exa
EmbeddingGemma Model Card	ai.google.dev official	2025-09-07	2026-10-07	hn
How to Fine-Tune Google's EmbeddingGemma – LLM	youtube.com	2025-09-06	2026-10-07	hn
EmbeddingGemma: The Best-in-Class Open Model for On-Device Embedding	developers.googleblog.com official	2025-09-04	2026-10-07	hn, exa
Welcome EmbeddingGemma, Google's new efficient embedding model	huggingface.co	2025-09-04	2026-10-07	hn, exa

Papers

1 paper

Papers 1

arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).

EmbeddingGemma: Powerful and Lightweight Text Representations this model

Henrique Schechter Vera, Sahil Dua, Biao Zhang +86 · 2025-09-24 · PDF · HF paper page (51 upvotes)

Similar models

3 models

Similar models

gte-multilingual-base · 305M
same size band (S), same task (sentence-similarity), permissive license
paraphrase-multilingual-mpnet-base-v2 · 278M
same size band (S), same task (sentence-similarity), fewer providers, permissive license
multilingual-e5-base · 278M
same size band (S), same task (sentence-similarity), fewer providers, permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:32 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 14:59 UTC · Last verified 2026-10-25

Follow embeddinggemma-300m

Watch this model

/feeds/models/google--embeddinggemma-300m.xml(RSS 2.0) · every change to embeddinggemma-300m
/feeds/families/google--embeddinggemma.xml(RSS 2.0) · every change across the embeddinggemma family

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.