TextText analysis

FRED-T5

Sber's Russian text-to-text model, successor to ruT5 (2021). Small and fast: fine-tuned for summarizing, paraphrasing and fixing errors in Russian text; ready-made SAGE spell-checking versions exist.

The last open version came out in Mar 2024. The family has not been updated for a long time: the model still works, but do not expect fixes or new sizes.

Developer
Sber (ai-forever), Russia
First release
Jan 2023
Latest release
Mar 2024
Sizes
95M – 1.7B
License
Commercial use allowedApache 2.0 (FRED-T5), MIT (spell-checking versions)
Russian
Supported
Running
On your own serverAlso runs without a GPU
Industries
Documents and accounting, Education, Customer support
Deploy on your server Hugging Face GitHub

What it does

  • Fixing spelling mistakes and typos in Russian text
  • Short summaries and paraphrasing
  • Normalizing requests and inquiries before processing
  • Base for low-cost narrow fine-tuning

Where it is used

Customer supportContent and editorial teamsProcessing requests and forms

Hardware requirements

LaptopLaptop or regular PC, up to 8 GB of VRAM — smaller versions
fits
1 GPUOne GPU with 16–80 GB — mid-size versions
no versions
ClusterServer with several GPUs — flagship versions
no versions

Versions

  1. SAGE FRED-T5 large и distilled 95M
  2. FRED-T5 large spell (орфография)
  3. FRED-T5 large (820M)
  4. FRED-T5 1.7B

How to run it

On your own serverWeights are downloaded from Hugging Face and served with vLLM (load and API) or llama.cpp (modest hardware). The model then runs in a closed network with no per-request fees.Hugging Face
How much hardware you needCalculate the VRAM for the model size, context length and number of concurrent requests.Open the hardware calculator

I can set this up end to end: pick the model size, deploy it on your server and connect it to your systems.

Frequently asked questions

Can FRED-T5 be used in a commercial project?

Yes. License: Apache 2.0 (FRED-T5), MIT (spell-checking versions). It allows commercial use, but it is still worth having a lawyer review the license before launch.

What hardware does FRED-T5 need?

At minimum: Laptop or regular PC, up to 8 GB of VRAM — smaller versions. Some versions also run on an ordinary CPU, without a GPU. You can calculate the exact VRAM for your model size and context in the hardware calculator.

Does FRED-T5 support Russian?

Yes, Russian is listed on the model card.

Where can I download FRED-T5 and what does it cost?

The FRED-T5 weights are open and free to download. You only pay for the hardware it runs on and for the setup. Source links are at the bottom of this page.

How I deploy it for clients

  1. SelectionI pick the model size for your task and hardware and test it on your examples.
  2. DeploymentI deploy it on your server or in a closed network and provide an API.
  3. Fine-tuningI fine-tune it on your data (LoRA) or connect a knowledge base — whichever is cheaper for the task.
  4. IntegrationI connect it to your CRM, ERP, bot, website or team chat and set up monitoring.

Similar models

TextFlan-T5 и Flan-UL2Google · USACommercial use allowed

Compact input-output models trained to follow instructions. Still used as a cheap base for classification, extraction and short answers.

DetailsText analysisruBERT, ruRoBERTa, ruELECTRA (ai-forever)SberDevices (ai-forever) · RussiaCommercial use allowed

Sber's Russian-language encoders trained on large Russian corpora. A base for classifiers, NER and semantic search in Russian.

DetailsTextruGPT-3.5Sber (ai-forever) · RussiaCommercial use allowed

Sber's 13-billion-parameter base Russian model; GigaChat grew out of its fine-tuned version. Continues texts in Russian and English, context only 2048 tokens; today useful as a base for narrow fine-tuning.

Details

Source: huggingface.co/ai-forever/sage-fredt5-large. Data checked against the model card on 22 Sep 2026. Have a lawyer review the license before commercial launch.