Artificial intelligence is on everyone's mind at the moment. The best-known models come almost exclusively from the USA. With Teuken-7B, Europe wants to show that it can help shape the field. 🎯
What makes Teuken-7B special?
Teuken-7B was developed as an alternative for research and for companies that value transparency, data protection and the ability to adapt a model themselves. It is based on an open-source approach – the code can be used freely and adapted commercially.
With seven billion parameters, Teuken-7B is smaller than GPT-4 (estimated at 200 billion parameters), but it shows an impressive range of languages and areas of application.
Ten partners from research and industry
The OpenGPT-X project is led by Fraunhofer IAIS and Fraunhofer IIS. Also involved are TU Dresden, Forschungszentrum Jülich, Aleph Alpha and IONOS SE.
“Our model has demonstrated its capability across a wide range of languages, and we hope that as many people as possible will adopt or develop it further.”
Stefan Wrobel, Fraunhofer IAIS
Why this matters for Europe
- Open source and freely available
- Usable in commercial projects
- Adaptable and transparent
- Data protection standards for the European market
