As part of the ongoing development of RWTHgpt, opportunities for API use in teaching and learning contexts are being further expanded.
For some time now, students—as well as staff—have been able to access selected RWTHgpt models using personal API keys. However, API usage was previously subject to certain volume restrictions.
As part of the Inferenz NRW initiative, which provides universities in North Rhine-Westphalia with a university-controlled infrastructure for the use of open-source language models, this access is now being significantly expanded: The previous limitations have been completely removed, so that the aforementioned models can be used freely under the regular terms of use.
As a result, API access is now available to students for the following models without additional usage restrictions:
- gpt-oss-120b
- e5-mistral-7b-instruct
- qwen3-embedding-8b
- mistralai-mistral-small-4-119b
The offering includes both powerful language models and specialized embedding models that can be used, for example, for Retrieval-Augmented Generation (RAG). The expansion specifically supports the use of RWTHgpt in courses, student projects, theses, and research applications. The goal is to further promote low-threshold access to modern open-source models and enable flexible use cases in the university environment.
Instructions for setting up an API key and for usage can be found on IT Center Help.
More information on new deployments can be found on our IT Center Changes Blog.