Qwen 3.5 family added to our secure AI ecosystem

This week, we have added a new model to EduGPT.

And to be honest… this is one that really makes a noticeable difference. We are now running on a model from the Qwen 3.5 family. With a context of up to 250k tokens and an output of up to 16k.

Simply put, this means you can continue working with AI along the same line of thought, even as things become larger and more complex.

But for us, the real advantage isn’t just in the power—it’s in the choices behind it.

We run this model entirely within our European infrastructure, via Scaleway.

- No training on user data.
- No unclear data flows.
- No dependence on parties that fall outside European legal jurisdiction.

AI only becomes truly interesting when it is not only intelligent, but also reliable, explainable, and under your control.

And it doesn’t have to cost millions.

This feels like a step in that direction.