Model reference · open weights

rubert-tiny-lite

Available as managed deployment Embeddings sergeyzh · community Embeddings 1 variants 591 dl/mo

rubert-tiny-lite is an open-weight embedding model from sergeyzh. AxForge deploys and operates it for you on dedicated EU-owned hardware — with the licence handled where one is required.

Available as managed deployment — configured and operated for you on dedicated EU hardware, quoted per deployment.

What it is

Released bysergeyzh
TypeEmbedding models
TaskEmbeddings
Parameters (lead)23M
Context512 tokens
Runs withsentence-transformers
Released2025-07-10
Popularity591 downloads / month
LicenceOpen weights

About

What rubert-tiny-lite is

Быстрая модель BERT для русского языка с размером ембеддинга 256 и длиной контекста 512. Модель получена методом последовательной дистилляции моделей sergeyzh/rubert-tiny-turbo и BAAI/bge-m3. Выигрывает по скорости у rubert-tiny-turbo при аналогичном качестве на CPU в ~x1.4, на GPU в ~x1.2 раза.

Read the full model card

Использование

from sentence_transformers import SentenceTransformer

model = SentenceTransformer('sergeyzh/rubert-tiny-lite')

sentences = ["привет мир", "hello world", "здравствуй вселенная"]
embeddings = model.encode(sentences)

print(model.similarity(embeddings, embeddings))

Метрики

Оценки модели на бенчмарке encodechka:

modelSTSPINLISATI
BAAI/bge-m30.8640.7490.5100.8190.973
intfloat/multilingual-e5-large0.8620.7270.4730.8100.979
sergeyzh/rubert-tiny-lite0.8390.7120.4880.7880.949
intfloat/multilingual-e5-base0.8350.7040.4590.7960.964
sergeyzh/rubert-tiny-turbo0.8280.7220.4760.7870.955
intfloat/multilingual-e5-small0.8220.7140.4570.7580.957
cointegrated/rubert-tiny20.7500.6510.4170.7370.937

Оценки модели на бенчмарке ruMTEB:

Model NameMetricrubert-tiny2rubert-tiny-turborubert-tiny-litemultilingual-e5-smallmultilingual-e5-basemultilingual-e5-large
CEDRClassificationAccuracy0.3690.3900.4070.4010.4230.448
GeoreviewClassificationAccuracy0.3960.4140.4230.4470.4610.497
GeoreviewClusteringP2PV-measure0.4420.5970.6110.5860.5450.605
HeadlineClassificationAccuracy0.7420.6860.6520.7320.7570.758
InappropriatenessClassificationAccuracy0.5860.5910.5880.5920.5880.616
KinopoiskClassificationAccuracy0.4910.5050.5070.5000.5090.566
RiaNewsRetrievalNDCG@100.1400.5130.6170.7000.7020.807
RuBQRerankingMAP@100.4610.6220.6310.7150.7200.756
RuBQRetrievalNDCG@100.1090.5170.5110.6850.6960.741
RuReviewsClassificationAccuracy0.5700.6070.6150.6120.6300.653
RuSTSBenchmarkSTSPearson correlation0.6940.7870.7990.7810.7960.831
RuSciBenchGRNTIClassificationAccuracy0.4560.5290.5440.5500.5630.582
RuSciBenchGRNTIClusteringP2PV-measure0.4140.4810.5100.5110.5160.520
RuSciBenchOECDClassificationAccuracy0.3550.4150.4240.4270.4230.445
RuSciBenchOECDClusteringP2PV-measure0.3810.4110.4380.4430.4480.450
SensitiveTopicsClassificationAccuracy0.2200.2440.2820.2280.2340.257
TERRaClassificationAverage Precision0.5190.5630.5740.5510.5500.584
Model NameMetricrubert-tiny2rubert-tiny-turborubert-tiny-litemultilingual-e5-smallmultilingual-e5-basemultilingual-e5-large
ClassificationAccuracy0.5140.535

From the published model card. Full card on the HuggingFace links in the sidebar.

Using it via the API

Call it like any OpenAI endpoint

Once AxForge deploys rubert-tiny-lite for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (rubert-tiny-lite below is illustrative; you get the exact model name on deployment.)

$ curl -sS https://api.axforge.ai/v1/embeddings \
  -H "Authorization: Bearer $AXFORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"rubert-tiny-lite","input":"text to embed"}'

Create an account — your API key is available in the console. 3M free tokens every 30 days with every new account.

© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms