The models run on the box.
Open-weight models such as Llama, Qwen and Gemma, running on the Exalt AI module inside the box. Which ones you run is decided with you.
From your material to a signed-off answer.
How the models are handled.
- Chosen
- In the consultation, against the work you want done and the material it answers from.
- Named
- By family. We don't publish versions or sizes here, because the lineup changes.
- Run
- On the box.
- Updates
- On a connected box, Exalt support applies new models and updates during a session your admin opens. If the box runs air-gapped, how they reach it is agreed for your deployment.
- Cited
- Answers point to the material they came from.
Put local AI to work where it matters most.
Tell us the work first. The models follow from it.