Meet Lux 1.0. Our own model, built in New Zealand.
Most AI companies rent someone else’s model. We build and tune our own — Lux 1.0 — so your data stays onshore, your costs stay predictable, and the model actually understands how businesses here work.
↳ Read how we built Lux 1.0The model
Lux 1.0, priced to scale.
A frontier-grade model you run on your terms — flat, predictable token pricing, a 1M-token context window, and your data kept onshore. No surprise vendor bills, no lock-in.
Context
1M
tokens per request
Residency
NZ
data stays onshore
Input
$0.25
per M tokens
Output
$0.50
per M tokens
What this is
Foundation models we own end-to-end.
We develop our own foundation models and fine-tune them on your domain — your documents, your workflows, your tone. Because we own the stack, we can deploy them privately, keep your data in New Zealand, and tune the model to your exact use case instead of forcing your problem into a generic chatbot.
Sovereign by default — your data is never sent to a third-party model provider.
How it helps
What you actually get.
- Trained and tuned in-house — not a thin wrapper over someone else’s API
- Your data stays onshore: private and sovereign deployment options
- Fine-tuned on your domain, your documents, and your tone of voice
- Runs where you need it — our cloud, your cloud, or on-prem
What’s included
- 01 Trained and tuned in-house — not a thin wrapper over someone else’s API
- 02 Your data stays onshore: private and sovereign deployment options
- 03 Fine-tuned on your domain, your documents, and your tone of voice
- 04 Runs where you need it — our cloud, your cloud, or on-prem
- 05 No per-token vendor lock-in and predictable, capped running costs
- 06 Full eval suite so you can see exactly how the model performs on your tasks
The essentials
Built for control, privacy, and fit.
- Deployment
- Managed cloud, your VPC, or fully on-prem
- Data residency
- New Zealand / Australia — your data never leaves
- Customisation
- Fine-tuning, RAG, and tool-use on your own data
- Integration
- Drop-in API, plus MCP and agent-framework support
Contact us
About our foundation models.
Curious whether your own tuned model makes sense? Tell us about your data and use case — we’ll reply within one NZ working day.