What is Inference markup?

Inference markup is when a platform charges more per token than the underlying model provider's own rate, effectively reselling AI usage at a margin.

Many AI products bundle model usage into their subscription and charge a margin on top of the provider's raw per-token price — often invisibly, since you never see the underlying provider's bill. This can make usage-heavy workloads unpredictable or expensive at scale.

A BYOK platform avoids this by having you hold the provider relationship directly: you're billed by the model provider at their published rate, and the platform charges separately (usually a flat fee) for hosting or orchestration.

On VibeOpenClaw

VibeOpenClaw charges a flat $24/mo (Pro) or $48/mo (Premium) for hosting and never marks up inference — every token you use is billed by your chosen provider directly, at their rate.