Wikier

GPT NTNU

This page contains information about the AI chat GPT NTNU (gpt.ntnu.no). This tool is available for students and staff at NTNU. GPT NTNU is developed in collaboration with UiO and runs on infrastructure at UiO and NTNU.

Norwegian version – GPT NTNU

Topic page about artificial intelligence in teaching and assessment | Pages tagged artificial intelligence

Get started

GPT NTNU is available on the website gpt.ntnu.no. Log in with your regular NTNU account by clicking the button “Log in via Feide”. Read and accept the terms of use before using the service.

  Login to GPT NTNU

There are several different language models available that you can choose from when starting a conversation. Read the descriptions and choose the one that best suits your needs.

During periods of high user volume, there may be delays and slow responses. Please try a different model or wait until a later time.

What models are available?

ModelApproved dataInfrastructureInput - best suited forDeveloped by
GPT-OSS 120 B🟢Open
🟡Internal
🔴Confidential*
Runs on NTNU computersText - powerful open-weight model for comprehensive analysis
*See Important when processing confidential data
OpenAI (USA)
NorwAI Magistral🟢Open
🟡Internal
Runs on NTNU computers -
still under development and testing
Text - Norwegian model from NTNU, complex reasoningNorwAI, NTNU (Norway)
GLM 5.2🟢Open
🟡Internal
🔴Confidential*
Runs on NTNU computersText - an open reasoning model, good for coding
*See Important when processing confidential data
Z.ai (China)
Kimi K2.6🟢Open
🟡Internal
Runs on NTNU computersText and image - one of the best new models, codingMoonshot AI (China)
Mistral Medium 3.5 128B🟢Open
🟡Internal
Runs on NTNU computersText and image - large european modelMistral AI SAS (France)
GPT-5.1 Thinking🟢Open
🟡Internal
Runs on Microsoft Azure cloud serviceText and image - powerful reasoning model for insightOpenAI (USA)

Several language models (AI that is made for reading and generating text) are available in GPT NTNU. The various models have different strengths and weaknesses, and continuous assessments will be made of which ones are most relevant. This is related to the type of input, what they are best suited for and how efficient and power-intensive they are. In addition, the models are trained on different types of data and can have different types of bias and filters, which will affect what you as a user get as output. It is important to be critical of what you get as output from the models. Consider ethical and legal aspects when using AI models, from energy use and training data; to security policy, privacy, bias and misinformation. Due to these challenges and the rapid development of models, a continuous assessments of which models should be available in GPT NTNU will be done.

Note: The models cannot search web pages on the internet, and only have data up to the point in time when they were trained. GPT NTNU will therefore not have up‑to‑date information, and is not a chatbot for user support or for referring users to websites.

Feel free to provide feedback and input on models and other development requests in this Nettskjema-form.

What can GPT NTNU be used for?

You can use the tool to process information and data that is classified as open (green) and internal (yellow). Examples of use include proofreading and improvement of texts, meeting minutes and summaries, explanation of academic and complex questions or review of key concepts in the syllabus before exams.

GPT NTNU as a service was in June 2026 assessed as secure enough for confidential (red) data, and two of the models are now approved for confidential (red) data without personal data. Examples of use are data that includes trade secrets and information subject to export control.

Important when processing confidential data

The IT department has conducted a technical assessment of the security level of GPT NTNU. As a general rule, AI tools should not be used to process personal data; specific assessments must be made and approved. GPT NTNU does not have a general approval for the use of personal data in the tool. GPT NTNU should not be used to assess, process or analyze information about individuals. Be careful with what personal data you enter, and feel free to use examples or anonymized data.

Always remember to delete chat history that contains confidential data when you are finished with processing data.

Administrative processes

GPT NTNU is not approved for the processing of personal data in administrative processes. The use of AI in administrative processes must be in accordance with the framework and requirements set by the process owner (cf. NTNU's guidelines for generative AI). The process owner is responsible for the necessary documentation and assessments, including the basis for processing, purpose and DPIA (data protection impact assessment).

Research and student projects

If you are planning on processing personal data (both internal and confidential) you must have a legal basis (see more about lawful basis). You must also consider possible risks and appropriateness of your specific use of AI. For advice and guidance on using GPT NTNU or other tools in research, contact Research Data @NTNU.

About GPT NTNU

GPT NTNU is a web application developed in collaboration with UiO, based on their service GPT UiO. The service runs on internal infrastructure at NTNU and UiO, and offers a GPT model from OpenAI, in addition to several other language models under testing.
GPT NTNU is a safe and controlled alternative that takes into account privacy for you as a user:

  • Conversations and user information are stored on a physical server at UiO
  • Data and conversations are not used to train the models further
  • The language models themselves run on infrastructure at NTNU (see more info on models available on the supercomputer IDUN)

User guide

More comprehensive guides are being developed. UiO has guides and detailed information about GPT UiO, and several of these can also provide useful tips for using GPT NTNU:

Always remember

  • Verify all information before use
  • Avoid pasting personal information directly into GPT NTNU
  • Clearly mark AI-generated content
  • You are responsible for everything you produce with AI tools like GPT NTNU
  • Start a new conversation for sensitive topics
  • Consider deleting chat history you no longer need

What is stored and for how long? - Privacy statement

When you use GPT NTNU, your queries (input) and answers (output) are stored on a server at UiO. You have access to this conversation history when you are logged in. When you delete a conversation, it will immediately be removed from your visible history. The conversation will be marked for permanent deletion from servers at UiO, and deleted at once for models running on NTNU computers and after one month for the model running on Azure. History from your inactive conversations will be automatically deleted after one year.

In addition, technical log data is stored for one year. This includes logins, time stamps, IP addresses and account information, in accordance with other IT services and tools at NTNU. This is to ensure the operation and maintenance of the service itself, including troubleshooting and quality, as well as to prevent fraud, abuse and maintain security.

To improve the user experience of the application itself in the browser, cookies are stored about login session, language selection and authentication.

NTNU has a data processing agreement with UiO that regulates the storage, access and use of personal data on their servers (see also the privacy statement for GPT UiO for further details, Norwegian only).

Logs and access to information

Only operational and security staff at NTNU and UiO have technical access to the logs. Access to these will only be granted after a thorough assessment, and all access is permanently logged. Approved purposes for access are security incidents, troubleshooting technical problems, legal obligations/orders from authorities or users' own requests for access.

Important: Chat history is not used for control purposes, to make decisions about individuals, or to investigate cheating. This is part of NTNU's guidelines. Users are informed about access either in advance or afterwards. In the event of security incidents or legal orders, the user is notified afterwards when access has been carried out.

Terms of Use for GPT NTNU

When you first log in to GPT NTNU, you must read and accept the terms of use for GPT NTNU, which are similar to Terms of Service at GPT UiO. In addition, all use must be in line with the NTNU ICT regulations, which apply to all employees, students and guests associated with NTNU.

Some particularly relevant points to be aware of when using an AI service such as GPT NTNU:

  • do not mislead third parties by claiming that output has been generated by humans when this is not the case
  • do not use the service to make decisions without appropriate human supervision, especially if the decision could have consequences that could affect individuals
  • do not use the service for content, including both input and output, that
    • describes, contains, promotes or is used for purposes related to child sexual abuse and sexual exploitation
    • describes or is used for purposes related to grooming
    • describes or is used for purposes related to human trafficking
    • describes, supports, promotes, glorifies, encourages and/or instructs individuals to commit suicide and self-harm
    • describes, contains or promotes violence, violent content and violent behaviour
    • describes, supports, promotes, glorifies, encourages and/or instructs individuals to commit terrorism and violent extremism
    • promotes or is used for purposes related to violent threats, incitement to or glorification of violence
    • is used for purposes related to hate speech and discrimination
    • used for purposes related to bullying and harassment
    • used for purposes related to intentional fraud, deception, and disinformation
    • used for inauthentic interactions, such as fake accounts, automated inauthentic activity, impersonation to obtain unauthorized information or privileges, and for content that claims to be from a person, company, government agency, or others without express permission

Learn More About Artificial Intelligence

Contact

If you experience technical problems, please contact the Orakel Support Service in NTNU Hjelp.

If you have feedback on this wiki page or curious about further development of GPT NTNU, please contact support for IDUN HPC in NTNU Help.

Child Pages (1)

  • Terms of use GPT NTNU

    Important when processing confidential (red) data: GPT NTNU as a service and two of the models (GLM and GPT-OSS) are considered secure enough for use on confidential (red) data. This is only a...