Qwen or GigaChat: which to choose for business

Both models handle Russian and scale from a regular computer to a multi-GPU cluster. Qwen covers a wider size range (0.6B to 2.4T-A95B) and ships in Ollama, but the larger Qwen3.8 models use their own license, so commercial use comes with conditions. GigaChat from Sber is fully MIT, built around the Russian language and local context, and the catalog lists it for banks, the public sector and companies that must keep data in Russia; it is not in Ollama. GigaChat was updated more recently (2026-09 vs 2026-08) and has an audio version that handles recordings up to two hours.

Comparison based on catalog data

ParameterQwenGigaChat
CategoryText, Image + textText, Voice assistants
DeveloperAlibaba, ChinaSber, Russia
ReleasesAug 2023 – Aug 2026Dec 2024 – Sep 2026
Sizes0,6B – 2,4T-A95B10B-A1.8B – 702B-A36B
HardwareLaptop, 1 GPU, ClusterLaptop, 1 GPU, Cluster
Commercial useCommercial use with conditionsCommercial use allowed
LicenseApache 2.0 (most versions); the larger Qwen3.8 models have their own licenseMIT
RussianSupportedSupported
OllamaYesNo
Without GPUYesYes
Tasks
  • Chatbot and knowledge-base assistant
  • Replies to emails and customer requests
  • Document parsing and classification
  • Product description generation
  • Russian-language employee assistant on your own server
  • Customer replies and request handling in Russian
  • Working with contracts and internal policies
  • Drafting emails and reports

Choose Qwen if

  • You need a very small model: Qwen starts at 0.6B, GigaChat starts at 10B-A1.8B
  • You want a one-command install through Ollama
  • Your tasks are e-commerce: product descriptions, ticket triage, knowledge base answers
Qwen

Choose GigaChat if

  • You want a simple license: every GigaChat version is MIT
  • You are a bank, a public body or a company required to keep data in Russia
  • You need audio: GigaChat3.1-Audio understands recordings up to two hours long
GigaChat

Other comparisons

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.

  1. SelectThe model and size for your task and hardware budget
  2. DeployOn your server or in a closed network, with an API
  3. Fine-tuneOn your data, or connect a knowledge base
  4. IntegrateInto your CRM, ERP, bot, website or team chat
Discuss deployment