[K]inference: Free Tokens for Everyone
November 19, 2026
Open-weight models are rapidly approaching (and in some cases even eclipsing) the quality of frontier models. And running them on Kubernetes is easier than you think. In this demo, Sarah and Flynn showcase [K]inference, Buoyant's latest open source project, and how it can help you build a cost-efficient, SLO-focused LLM deployment as part of your Kubernetes platform using open-weight models and commonly available cloud hardware, all without breaking a sweat. They cover metrics to monitor, inference types, the benefits of KV cache-aware routing, and how to evaluate the economics of your LLM deployment— and make your CFO happy.