Models and cost

Near-frontier models. Open-source prices.

NIK Agent defaults to our own private model, running open-source weights in infrastructure we host domestically. A frontier model is one setting away if you want one, with your key and your rates, and it never becomes the default behind your back.

The default

Our private model, and why it is the default

Open-source weights have closed most of the practical gap for the work an assistant actually does: reading, summarizing, drafting, extracting and reasoning over your own material. Running them ourselves is what makes sovereignty and the price both possible.

Hosted domestically

Inference runs in infrastructure we operate in the United States, so your prompts, documents and traces stay in country.

Never trained on your data

Your conversations are not used to train a model and are not pooled with other customers. They are governed by your organization's retention policy.

Metered, not bundled

$0.25 per million tokens with no minimum, so a light user costs cents and a heavy user is visible before the invoice arrives.

Chain of thought inside

Reasoning is part of the product rather than a vendor feature you hope stays enabled, so the trace is always there for the user and the administrator.

Sized for the task

Fast models for quick questions and larger models for long documents, chosen by the agent rather than by you.

Independent of one lab

Open weights mean the roadmap is not hostage to a single vendor's pricing page, deprecations or terms of service.
Bring your own frontier model

Frontier capability on your terms

If you want Claude, GPT, Gemini or Grok in the loop, an administrator adds the organization's API key and chooses who may use it. The requests bill to your account with that vendor, and the rest of the workspace stays on the private model.

  • One key, per organization, stored for your workspace and never shared across customers.
  • Per group enablement, so finance can have it while the rest of the company keeps the private model.
  • The same interface and the same chain-of-thought view regardless of which model answered.
  • Vendor terms apply to those requests. We tell you that plainly instead of implying otherwise.
  • Switch a group back to the private model at any time, with no reinstall.
What changes when you add a frontier key
Data locationThose requests go to the vendor's cloud
TrainingGoverned by that vendor's policy
CostBilled by the vendor at their rates
Everything elseUnchanged: identity, guardrails, traces, admin control
Cost in practice

Why the number looks small next to a frontier seat

A per-seat frontier plan bundles inference into a flat price that has to cover its heaviest users. Metering the tokens and paying $5 for the seat keeps the predictable part predictable and the variable part honest.

ComponentWhat you payWho sees it
Seat$5 per user per month on Team, $0 on FreeFinance, once a month
Private model tokens$0.25 per million tokensAdministrators, by person and by team
Frontier model tokensThe vendor's rates, on your own keyYou, in the vendor's console
Free plan20 million tokens a week, refreshed weeklyNobody, it is free

For the wider Netarx platform, including the identity layer that NIK Agent builds on, see netarx.com pricing.

Try it on the private model first

Most teams find the default is enough. Send a document the agent has never seen and judge the answer, not the benchmark.