DeepSeek, a Chinese AI startup backed by High-Flyer, said on October 20 that it had open-sourced DeepInfer, an inference framework for its DeepSeek-V4.1-Flash model.

It is designed to reduce deployment costs on Nvidia’s SM100 Blackwell graphics processing units and Huawei’s Ascend 950 chips.

Inference is the process of serving and running AI models, and frameworks such as DeepInfer are used to optimize that step.

DeepSeek introduced DeepSeek-V4.1-Flash on September 9, 2026, as a separate model variant.

Its application programming interface change log shows Flash as part of the company’s production model lineup.

DeepInfer is publicly available in DeepSeek’s GitHub organization, where the company describes it as software for high-performance, large-scale inference.

🔗 Source: South China Morning Post

Recent DeepSeek developments