DeepSeek is open-sourcing DeepInfer, the vLLM-based inference engine behind its fast and cheap V4.1-Flash model, which cuts compute costs and energy use.

Deepseek has open-sourced DeepInfer, its internal inference framework, under the Apache License 2.0. Inference is how a trained AI model handles requests and generates responses. DeepInfer is the serving software that lets Deepseek's V4.1-Flash model produce answers quickly at scale.

The framework builds on vLLM and adds optimized compute kernels, distributed processing across multiple chips, and faster text generation techniques. Code for future Deepseek models will also ship as open source. Deepseek is additionally releasing deepseek-recipe, which replicates the token processing of the official Deepseek API so that self-hosted setups behave the same way.

Deepseek's models are already very inexpensive to use, and DeepInfer addresses the remaining gap in running them more cheaply. The company said on X that its goal is "to help the community lower compute costs, reduce energy use, and make AI more affordable."

The release also adds pricing pressure on Western providers such as OpenAI and Anthropic. Given the AI race between China and the United States, the move could be politically motivated as well. If Western AI companies miss their growth targets, that would pose a risk to the US economy and beyond.