Demo

Talk to a model on our stack.

An instruction-tuned model served from our own inference setup on 1× RTX 5090. It answers from this site's published material and nothing else.

Connecting to the model… 750M-parameter GLM · 1× RTX 5090 Demo unavailable

Starting up. Ask about the architecture. Not available right now.

This demo runs on hardware in our London office and is not publicly available.

Responses are generated by the model and may be imprecise.

Verification

One check before we start.

The model runs on a single machine in our London office. This keeps automated traffic off it, so there is capacity left for you.