Sinun AI
fi

Models

Qwen 3.5 — the open-source coding king

Qwen 3.5 27B (Apache 2.0) hits 72.4% on SWE-Bench Verified and fits on a single 16 GB GPU. Paired with Viking or Poro it makes a strong dual setup: Finnish + code.

Apache 2.072.4% SWE-Bench16 GB is enough

For developers, data scientists and researchers, Qwen 3.5 is the top local pick for coding and reasoning. Apache 2.0 allows royalty-free commercial use.

Where Qwen 3.5 shines

Writing and refactoring code, SQL queries, data-analysis scripts, mathematical reasoning, software design. One 16 GB GPU is enough — many researchers and entrepreneurs don't need more.

Pairing with Viking

Sinun AI can route queries automatically: a Finnish conversation to Viking, a technical coding query to Qwen. This dual setup gives you linguistic and technical top quality without a cloud.

Frequently asked

Is Qwen risky due to country of origin?
An open-weight model running on your hardware makes no calls back to the maker. The weights can be audited; data never leaves your device. Risks are usually less about the provider and more about where inference runs.

Updated 2026-04-21

Related

Models

Viking — the Nordic open LLM from Kajaani

Viking 7B/13B/33B is Silo AI's open LLM for Finnish, Swedish, Norwegian, Danish and Icelandic. Apache 2.0, trained on the LUMI supercomputer in Kajaani.

Models

Poro — Europe's trailblazer open language model

Poro is Silo AI's and TurkuNLP's open LLM for Finnish, English and code. Trained on LUMI in Kajaani, the predecessor of Viking.

Models

Gemma 4 — the best open edge model of 2026

Gemma 4 and Gemma 4 E4B are Google's open multimodal models. E4B runs in 3 GB VRAM, and the full Gemma 4 fits on a mini-PC or Jetson edge device.

Models

Mistral Small 4 — light, European, multilingual

Mistral Small 4 is Mistral AI's (Paris) open model — light to run, fluent across European languages. Built and trained in the EU.

All resources

Want your own local AI assistant?

Tell us about your work and hardware — we'll map the right model, the right hardware tier and the right sync configuration.

Get in Touch