Saiga or Vikhr: which Russian fine-tune to choose
Both are Russian-language fine-tunes of other people's open models, and both are installed by hand since neither is in Ollama. Saiga inherits its license from the base model: the LoRA adapters are CC-BY 4.0, but the assembled model follows the rules of Llama, Gemma, Mistral Nemo or YandexGPT, so commercial use has to be checked per build. Vikhr keeps its main versions under Apache 2.0, starts at 0.5B and runs on a CPU, and Borealis recognizes and understands Russian speech. Vikhr was updated later, December 2025 against April 2025 for Saiga, while Saiga reaches up to 70B.
Comparison based on catalog data
| Parameter | Saiga | Vikhr |
|---|---|---|
| Category | Text | Text, Speech to text, Voice assistants |
| Developer | Ilya Gusev (IlyaGusev), Russia | Vikhr Models, Russia |
| Releases | Apr 2023 – Apr 2025 | Jan 2024 – Dec 2025 |
| Sizes | 7B – 70B | 0.5B – 24B |
| Hardware | Laptop, 1 GPU | Laptop, 1 GPU |
| Commercial use | Commercial use with conditions | Commercial use allowed |
| License | LoRA adapters: CC-BY 4.0; the complete model inherits the base license: Llama Community License, Gemma Terms, Apache 2.0 for the Mistral Nemo version, YandexGPT's own license for the YandexGPT 5 Lite version | Main versions: Apache 2.0; some fine-tunes inherit the base model license (Llama, YandexGPT) |
| Russian | Supported | Supported |
| Ollama | No | No |
| Without GPU | No | Yes |
| Tasks |
|
|
Choose Saiga if
- You need up to 70B for an assistant on your own server
- You already build on Llama, Gemma or Mistral Nemo
- Support, marketing and content work in Russian
Choose Vikhr if
- You want straightforward Apache 2.0 on the main versions
- No GPU available: Vikhr starts at 0.5B and runs on a CPU
- You need speech: Borealis turns Russian speech into text
Other comparisons
- Qwen or GigaChat: which to choose for business
- Qwen or Llama: which to choose for business
- GigaChat or YandexGPT: which to choose for business
- DeepSeek or Qwen: which to choose for business
- Gemma or Llama: which to choose for business
- Mistral or Llama: which to choose for business
- Whisper or GigaAM: which to choose for business
- Whisper or Parakeet: which to choose for business
- gpt-oss or Qwen: which one to run on your own server
- Llama or DeepSeek: which to run inside your own perimeter
- Mistral or Qwen: which to choose for business
- Gemma or Phi: small models for modest hardware
Need a model for your task?
An open model can run on your own server: data stays in-house, there is no per-request fee, and the model can be fine-tuned on your documents.
- SelectThe model and size for your task and hardware budget
- DeployOn your server or in a closed network, with an API
- Fine-tuneOn your data, or connect a knowledge base
- IntegrateInto your CRM, ERP, bot, website or team chat