DeepSeek Open-Sources DeepInfer, Its AI Inference Engine, for Nvidia and Huawei Chips

DeepSeek logo

Chinese AI lab DeepSeek has open-sourced DeepInfer, the inference framework it says powers its own large-scale production AI services, publishing the code on GitHub on October 20 under the permissive Apache 2.0 license. The release came just a day after the US government ordered its agencies to remove DeepSeek's models from federal systems.

DeepInfer is the software that takes a trained AI model and serves it to users at scale — splitting work across chips, batching requests and deciding how fast answers come back. The first release supports a single model, DeepSeek-V4.1-Flash, and runs on two very different chip families: Nvidia's newest SM100 Blackwell GPUs and Huawei's Ascend 950 neural processing units. No other frontier AI lab has treated a Chinese chip as an equal to Nvidia's in a serving framework.

DeepSeek also published rare production numbers: a single DeepInfer cluster serves 155 trillion tokens a day with a 96% input cache hit rate, earning about $38 of revenue per GPU per hour at off-peak prices. The code is built mostly on the open-source vLLM engine and bundles techniques like DSpark speculative decoding, which drafts several tokens ahead and verifies them in one pass for 4–7% higher throughput.

Source: read more

Comments

Popular posts from this blog

Fairbairn's Walk-Off Field Goal Seals Wembley Thriller as Texans Edge Jaguars 34-31

Anthropic IPO: Public S-1 Filing Expected Today Ahead of $2 Trillion Listing

Netflix Beats Earnings but Slashes Ad Forecast — Stock Tumbles After Hours