On Oct. 20, Chinese large-model company DeepSeek announced it is fully open-sourcing its internal inference framework, DeepInfer. DeepSeek said it hopes to share its in-house deployment experience to reduce computing costs and energy consumption, making intelligence accessible at lower prices to a broader audience.
Optimized for the model inference and deployment stage, the open-source framework is designed to help smaller vendors and developers improve inference efficiency and compress token-call costs — another significant move by a Chinese large-model company in the open-source ecosystem.
