Qwen3.8-27B: an Apache 2.0 Open Model You Can Deploy Inside Your Perimeter

In short. Qwen3.8-27B is a 27-billion-parameter open model from the Qwen team, published under the Apache 2.0 licence. It understands text, images and video, natively handles a 262,144-token context extensible to 1 million, and deploys with SGLang and vLLM. For companies that cannot send data to external clouds, it is another strong candidate for an AI assistant inside their own perimeter.

What we know about the model

• 27 billion parameters, Apache 2.0 licence.
• Understanding of text, images and video.
• A native context of 262,144 tokens, extensible to 1,000,000.
• Recommended serving engines are SGLang and vLLM; an FP8 variant and quantised versions for llama.cpp, Ollama and LM Studio are available.

What it means for companies in Kazakhstan

The long context lets the model take regulations, contracts and technical documentation in full, and image understanding helps with scans and photos of documents. An open licence and deployment on your own server remove the main barrier for banks, manufacturers and the public sector: data never leaves the company's perimeter. Quality in Kazakh should be tested on your own tasks, as the model card does not list supported languages.

How 105 works with such models

105 deploys open language models inside the customer's perimeter: enterprise assistants and document search that cites its sources. More in the Enterprise LLMs section.