How to deploy and self-host DeepSeek-V3.1 on Northflank

DeepSeek-V3.1 is a significant advancement in the DeepSeek family of large language models, offering a 671B parameter Mixture-of-Experts architecture with a 128K context window for enhanced reasoning capabilities. This model supports both chat and think inference modes, allowing users to toggle between standard interactions and more reasoning-intensive tasks via an Open WebUI interface. It operates efficiently on 8× NVIDIA H200 GPUs using vLLM, providing high-throughput inference and improved reasoning speed compared to its predecessors. Users can deploy or self-host DeepSeek-V3.1 on the Northflank platform using either a one-click template or a manual setup, ensuring flexibility in deployment while avoiding rate limits with an OpenAI-compatible API. The model's open-weight design allows for secure, scalable deployment with cost-efficient pricing, making it one of the most capable open-weight language models available.