ROLEPL-AI publishes its LLM under an open license

jzaraik
Aug 27
2 min read

We are proud to announce that the ROLEPL-AI project has reached a key milestone: the release of its Large Language Model (LLM), trained specifically for role-playing simulations in education and vocational training.

Why a dedicated LLM?

Throughout the project, our consortium explored how artificial intelligence could support the development of soft skills through immersive simulations. Using pedagogical content created by our partner institutions, we trained a state-of-the-art language model to interact with learners in realistic role-play scenarios. The model is designed to provide natural, context-aware conversations, as well as constructive feedback on communication, problem-solving, and conflict management.

What makes this model unique?

Unlike general-purpose LLMs, the ROLEPL-AI model was fine-tuned with educational objectives in mind. It integrates scenarios, conflicts, and evaluation criteria developed by teachers and adapted to two very different learner groups: higher education students and adults distant from employment. This dual focus makes the model flexible and relevant for diverse training contexts.

Open and reusable

In line with Erasmus+ values, the ROLEPL-AI LLM is made available under a permissive open license. This means researchers, educators, and developers can reuse, adapt, or extend the model for their own projects. Whether for testing new pedagogical approaches, integrating into virtual environments, or conducting further research, this resource is now accessible to the wider community.

What’s next?

The publication of the LLM is not an end point, but a foundation. We hope that other institutions and innovators will build on this work to further explore the role of AI in education. By sharing our results openly, we aim to foster collaboration, inspire new applications, and contribute to responsible and impactful use of AI in learning.

You can access the LLM and its documentation here: