DeepSeek Dials Up the Pressure on OpenAI and Anthropic With Open-Source Inference Tool

DeepSeek has released DeepInfer, an open-source inference engine that allows developers to run a 67 B model on a single-node GPU cluster. This tool challenges proprietary stacks like OpenAI's hosted API and NVIDIA's TensorRT-LLM. Available via GitHub with an MIT-style license, Docker images on Hugging Face, and a Python toolkit, DeepInfer provides the same low-level kernels used internally. This release may attract more American businesses toward Chinese open-source models as benchmark data circulates online.