EmbeddingGemma 300M

Sep 2025
EmbeddingGemma 300M
Compare in
on
$0.002per 1M tokens
55.7MTEB English v2
Sep 2025released
2K tokensMax input

Rankingconfiguration unstated?

Full ranking →
1Harrier OSS v1 27B74.32KaLM Embedding Gemma3 12B 251172.33Qwen 3 Embedding 8B70.64Seed1.6 Embedding 121570.35Llama Embed Nemotron 8B69.5
···
29Text Multilingual Embedding 00262.230EmbeddingGemma 300M61.131GritLM 7B60.9

Strong at

B
Embeddings & Retrieval#53 of 200

Alternatives?

ModelScoreCostPriceWhat changes
BetterQwen 3 Embedding 8BunstatedOpen+15%+400%$0.01Memory (4-bit): 0.2 GB → 4.5 GBOpenKaLM Embedding Gemma3 12B 2511+18%——Memory (4-bit): 0.2 GB → 7.1 GBBetterQwen 3 Embedding 0.6BunstatedOpen+5%+400%$0.01Memory (4-bit): 0.2 GB → 0.4 GBiBetterMultilingual E5 Large InstructunstatedOpen+3%+400%$0.01Max input: 2K tokens → 1K tokens+1BetterQwen 3 Embedding 4BunstatedOpen+14%+900%$0.02Memory (4-bit): 0.2 GB → 2.4 GB

Run it yourself

estimate
Fits on✓ 8 GB (phone, laptop) · full quality✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceGemma termscommercial use with conditions

Benchmarks

in Embeddings & Retrieval
MTEB Persian v2 · Mean over tasks · Persian67.8#8
MTEB Persian v2 · Mean over task types · Persian70.8#8
MTEB Persian v2 · STS · Persian79.1#10
MTEB Persian v2 · Clustering · Persian63.7#12
MTEB Persian v2 · Classification · Persian68.7#11
MTEB Persian v2 · Retrieval · Persian59.2#5
MTEB Persian v2 · Reranking · Persian71.5#15
MTEB Persian v2 · Bitext Mining · Persian69.5#8
MTEB Persian v2 · Pair Classification · Persian84.2#12
JMTEB v2 · Clustering · Japanese53.1#20
JMTEB v2 · Classification · Japanese71.6#8
JMTEB v2 · STS · Japanese84.7#6
MTEB Europe v1 · Bitext Mining · multilingual (28 languages)80.8#77
MTEB Europe v1 · Clustering · multilingual (29 languages)46.5#32
MTEB Europe v1 · Instruction Reranking · English5.6#8
MTEB Europe v1 · Reranking · multilingual (12 languages)60#32
MTEB Europe v1 · Multilabel Classification · multilingual (23 languages)18.7#36
MTEB French v1 · Pair Classification · French62#52
MTEB Indic v1 · Reranking · multilingual (2 languages)88.8#4
MTEB Indic v1 · Pair Classification · Hindi81.1#14
MTEB Indic v1 · STS · multilingual (13 languages)43.1#44
MTEB Indic v1 · Clustering · multilingual (21 languages)35.2#46
MTEB Dutch v1 · Mean over tasks · Dutch57.6#21
MTEB Dutch v1 · Mean over task types · Dutch63.8#21
MTEB Dutch v1 · Classification · Dutch64.3#23
MTEB Dutch v1 · Clustering · Dutch34.2#37
MTEB Dutch v1 · Retrieval · Dutch61.7#20
MTEB Dutch v1 · STS · Dutch80.1#25
MTEB Dutch v1 · Multilabel Classification · Dutch37.3#23
MTEB Dutch v1 · Pair Classification · Dutch78.3#29
MTEB Dutch v1 · Reranking · Dutch90.5#27
MTEB Portuguese v1 · Mean over tasks · Portuguese73#1
MTEB Portuguese v1 · Mean over task types · Portuguese66.6#1
MTEB Portuguese v1 · Classification · Portuguese66.5#2
MTEB Portuguese v1 · Retrieval · Portuguese78.5#2
MTEB Portuguese v1 · Multilabel Classification · Portuguese21.5#44
MTEB Portuguese v1 · STS · Portuguese81.1#7
MTEB Portuguese v1 · Reranking · Portuguese85.3#3
MTEB Medical v1 · Clustering · English43#33
FollowIR · Mean over tasks · English5.6#8
FollowIR · Mean over task types · English5.6#8
FollowIR · Instruction Reranking · English5.6#8
MTEB Spanish v1 · Pair Classification · Spanish72.7#19
MTEB Thai v1 · Clustering · Thai46.8#46
MTEB Thai v1 · Pair Classification · Thai81.5#13
MTEB Scandinavian v1 · Bitext Mining · multilingual (3 languages)62.7#110
CoIR · Mean over tasks · multilingual (9 languages)67.3#37
CoIR · Mean over task types · multilingual (9 languages)67.3#37
CoIR · Retrieval · multilingual (9 languages)67.3#37
MTEB Code v1 · Mean over tasks · multilingual (15 languages)68.8#31
MTEB Code v1 · Mean over task types · multilingual (15 languages)68.8#31
MTEB Code v1 · Retrieval · multilingual (15 languages)68.8#31
MTEB Russian v1.1 · Mean over tasks · Russian65.2#15
MTEB Russian v1.1 · Mean over task types · Russian63.2#14
MTEB Russian v1.1 · Classification · Russian67.9#23
MTEB Russian v1.1 · STS · Russian73.7#23
MTEB Russian v1.1 · Reranking · Russian63.4#40
MTEB Russian v1.1 · Clustering · Russian59.7#25
MTEB Russian v1.1 · Pair Classification · Russian65.1#17
MTEB Russian v1.1 · Retrieval · Russian71.5#17
MTEB Russian v1.1 · Multilabel Classification · Russian40.6#25
MTEB multilingual v2 · Mean over tasks · multilingual (1037 languages)61.1#30
MTEB multilingual v2 · Mean over task types · multilingual (1037 languages)54.3#29
MTEB multilingual v2 · Classification · multilingual (119 languages)60.9#37
MTEB multilingual v2 · Clustering · multilingual (208 languages)51.2#32
MTEB multilingual v2 · STS · multilingual (35 languages)74.7#24
MTEB multilingual v2 · Pair Classification · multilingual (24 languages)81.4#19
MTEB multilingual v2 · Retrieval · multilingual (120 languages)62.5#31
MTEB multilingual v2 · Bitext Mining · multilingual (1026 languages)64.4#58
MTEB multilingual v2 · Instruction Reranking · English5.6#8
MTEB multilingual v2 · Multilabel Classification · multilingual (25 languages)24.7#30
MTEB multilingual v2 · Reranking · multilingual (22 languages)63.2#40

FAQ

How much does EmbeddingGemma 300M cost?›

$0.002 per 1M tokens.

Is EmbeddingGemma 300M free to use?›

There is no free API; the cheapest paid price is $0.002 per 1M input tokens. The weights are open, so you can run it yourself.

How good is EmbeddingGemma 300M, and what is it best at?›

#53 of 200 in Embeddings.

Is EmbeddingGemma 300M better than Text Multilingual Embedding 002?›

On MTEB multilingual v2 · Mean over tasks · multilingual (1037 languages), EmbeddingGemma 300M scores 61.1 vs 62.2 — lower. It is 92% cheaper.

How can I use EmbeddingGemma 300M?›

Through the Google API.

When was EmbeddingGemma 300M released? Is there a newer version?›

Released 2025-09-04. It is the current version in its line.

Is EmbeddingGemma 300M open source?›

Yes, the weights are open (gemma licence).

Can I run EmbeddingGemma 300M locally?›

The weights are open and can be self-hosted; hardware needs are not confirmed yet.

What is EmbeddingGemma 300M?›

EmbeddingGemma 300M is an open-weights embeddings & retrieval model from Google, released Sep 2025.

What are the best alternatives to EmbeddingGemma 300M?›

Qwen 3 Embedding 8B (better), KaLM Embedding Gemma3 12B 2511 (open weights).