Leveling up our NebulaONE model catalogue- what's new

20 Oct 2026

Our NebulaONE model catalogue has had an upgrade, thanks to collaborative team efforts

Cross collaboration work within IT Services from our AI team including David Kennedy and Fatima Mehmood, and the wider Digital Workplace team, has resulted in the expansion of our NebulaONE model catalogue with Kimi, Mistral and Llama, along withthe latest OpenAI models, GPT-5.6 Luna and Terra.

We have also successfully migrated existing agents from older OpenAI models to newer versions, giving users access to more up to date knowledge, improved processing and reduced operating costs.

This expansion gives users within the University a greater choice, recognising that different models and providers have different strengths across reasoning, coding and general purpose tasks. NebulaOne provides a supported and governed environment in which users can experiment and build with these models, while the University retains control over data boundaries, model availability, cost and lifecycle management. Usage of these new models remains within our Azure and NebulaOne control boundary, offering a more governed alternative to users accessing these model providers directly.

Looking forwards, we will soon be gathering feedback on the value of these new models, particularly for coding and other specialist use cases. This feedback will be valuable, along with user reqesusts, to improving this catalogue further and assessing the addition of other widely used models, including those offered by vendors such as Anthropic.