Kontinent
Kontinent provides an AI model gateway, dedicated inference infrastructure and model training services for businesses. · Infrastructure
How they use Nebius Token Factory
Kontinent uses Nebius Token Factory shared inference endpoints through its AI gateway. It includes only EU-hosted Nebius models in its catalogue. Kontinent notes that the provider API’s country field for shared endpoints is advisory and may change. Customer requests reach individual inference providers according to customer configuration and permitted routing rules.
Kontinent also sends its own probe traffic through the gateway to measure model availability. Its status page includes Nebius Token Factory models. The measurement routine sends the same short prompt every 30 minutes and a longer prompt once a day. These measurements use Kontinent’s probes rather than customer traffic. When a fallback provider answers, the page identifies that provider.
Kontinent states that zero data retention and a training opt-out are enabled for its Nebius account. It distinguishes these account settings from contractual guarantees.
Sources
3 sourcesKontinent Nebius Token Factory disclosure
18 Oct 2026 · Kontinent · Web page
Kontinent uses Nebius Token Factory shared inference endpoints and carries only EU-hosted models in its catalogue. It states that the provider API’s country field for shared endpoints is advisory and may change. Kontinent also reports enabling zero data retention for its account.
Open original sourceKontinent measured model availability
18 Oct 2026 · Kontinent · Web page
Kontinent’s status page lists availability and latency measurements for Nebius Token Factory models and identifies calls served by other providers through fallback chains.
Open original sourceCompany context
kontinent.aiKontinent, operated by Berlin-based Synaptix GmbH, provides AI infrastructure for businesses and engineering teams that need European data processing. Its LLM Gateway offers an OpenAI-compatible API for accessing models across European providers, with routing, failover, consolidated billing and governance controls. Dedicated inference runs customers’ trained models on reserved GPUs in the EU. Its training services cover fine-tuning, distillation and reinforcement learning, supported by evaluation tools, checkpoint management and deployment monitoring. Customers can export their model weights, evaluations and documentation.
Kontinent official website and legal notice