SI Fly← All launches · #032 · 2026-10-29
DeepInfer
DeepSeek's production inference engine, open source
DeepInfer is DeepSeek's open-source inference framework that powers its large-scale production services. This release does offline inference for DeepSeek-V4.1-Flash only, on NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs. A KV cache system, prefill/decode disaggregation and cross-replica load balancing are not included. Apache-2.0, built primarily on vLLM.
Share this launch
Download PNG (transparent) · Download SVG
Embed this badge on your site (links back to this page):
<a href="https://siflie.com/launches/deepinfer/"><img src="https://siflie.com/badges/featured-on-si-fly.png" alt="Featured on SI Fly" width="220" height="44"></a>
Markdown:
[](https://siflie.com/launches/deepinfer/)
Plain text:
Featured on SI Fly — DeepInfer is launch #032: https://siflie.com/launches/deepinfer/
This badge means DeepInfer was featured on SI Fly as launch #032. It is not an endorsement or certification of the tool's quality, safety, or business performance. Use it only while this launch page is live; if the tool is delisted, please remove the badge.
Launch pages describe tools based on public information provided by their makers. SI Fly does not endorse tools and makes no claims about their performance.