↓ Skip to main content
fremai

Pricing · Coming soon

One plan, every model

€30 per seat, all models and features included - no capability locked behind price. fremai is prepaid: each seat bundles €10/month of usage pooled across your org, and you top up beyond it, paying only for the tokens you use. Served on EU-sovereign infrastructure and never trained on your prompts.

Plans

One standard plan; you pick a plan, not a capability - every plan includes all models, governance, SSO and audit export. Plans differ on seats and contracting, never on features.

  • fremai plan - €30 / seat / month - all models and features. €25.50/seat on annual billing (−15%). Prepaid: each seat bundles €10/month of usage, pooled org-wide and reset monthly.
  • Prepaid, pay-as-you-go beyond the credit - hold a credit balance; usage draws it down at each model’s published per-token rate, and you top up when you like. No auto-renew surprises.
  • 30-day free trial - one seat, all features and a starting credit, free for 30 days (card verification to prevent abuse). No permanent free tier.
  • Already on fremforge? - add fremai to your fremforge subscription at €20 / seat - one bill, one DPA.
  • Pay by card - card and common European payment methods; you get a VAT receipt in EUR (reverse-charge for EU businesses). Billed by fremverk, a European company. (Full e-invoicing is an Enterprise option - below.)
  • Budgets + alerts - spend controls per key, team or user with budgets, forecasts and 50 / 80 / 100% alerts, so a runaway job can’t surprise you.

Larger or regulated buyer? Enterprise adds prepaid volume rates, EAN/OIOUBL e-invoicing (NemHandel/Peppol), DKK invoicing, a negotiated DPA, SLA and named onboarding. Talk to us →

Launch line-up

Every model is included on every plan; usage is metered per-token against your balance. Exact context windows and per-token rates are published as each model is provisioned - here and in the /models API response.

GLM

GLM 5.2

Zhipu AI's flagship general + coding model. Strong tool-use and long-context reasoning.

Coming soon

M3

MiniMax M3

MiniMax's latest high-throughput model - built for fast, cost-efficient agentic workloads.

Coming soon

V4

DeepSeek V4

DeepSeek's frontier reasoning + coding model. A strong default for hard technical tasks.

Coming soon

K2

Kimi 2.7

Moonshot's Kimi - long-context specialist for large documents and codebases.

Coming soon

The catalog grows beyond the launch line-up over time. Every model is referenced by name, so moving one between backends is invisible to your code.

One endpoint, every model

The catalog above is reachable through a single OpenAI-compatible API. Nothing new to learn - if your code already talks to OpenAI, it already talks to fremai.

  • Drop-in OpenAI-compatible - point your SDK at https://api.fremai.eu/v1 and keep your code. Works with the OpenAI SDKs, curl, LangChain, opencode - anything that speaks the OpenAI API.
  • Switch by name - the whole catalog lives behind one endpoint. Change the model field to move between GLM, DeepSeek, Kimi and the rest; no new URLs, no new keys.
  • EU-sovereign - inference on EU infrastructure, control plane on Deutsche Telekom T Cloud in Germany. Your data stays in Europe.
  • Metadata-only - prompts and completions are never stored or trained on. We meter usage, not your content.

Questions on models or pricing?

Email hello@fremai.eu - happy to walk through it.