DeepSeek's October 2026 open-source release of DeepInfer left out a cache system, the split between prompt processing and answer generation, and load balancing across machines

On October 20, 2026 DeepSeek open-sourced DeepInfer, the in-house inference framework it uses to serve its models, under the Apache 2.0 license; at release it supported only DeepSeek-V4.1-Flash on Nvidia SM100 GPUs and Huawei Ascend 950 chips.