Optimizing GPT-OSS on NVIDIA DGX Spark: enabling local large model deployment
Original title:Tech Blog Optimizing GPT-OSS on NVIDIA DGX Spark: Getting the Most Out of Your Spark We’ve got some exciting updates about the NVIDIA DGX Spark\! In the week following the official launch, we collaborated closely with NVIDIA and successfully brought GPT-OSS 20B and GPT-OSS 120B suppor... Jerry Zhou November 3, 2025
AI summary
In partnership with NVIDIA, GPT-OSS 20B and 120B models have been successfully supported on DGX Spark via SGLang, achieving generation speeds of about 70 tokens/s for the 20B version and about 50 tokens/s for the 120B version, the best performance to date. Users can deploy the SGLang service via Docker and connect it to Open WebUI for local chat, or use LMRouter to convert request formats and run Claude Code fully locally. The solution makes it possible to deploy local coding agents with tens of billions of parameters on DGX Spark.
Key Takeaways
- In partnership with NVIDIA,GPT-OSS 20B and 120B models have been successfully supported on DGX…
- Users can deploy the SGLang service via Docker and connect it to Open WebUI for local chat,or use…
- The solution makes it possible to deploy local coding agents with tens of billions of parameters…
Quick Facts
- Published
- Updated
- Source
- LMSYS Blog
- Category
- Other
DGX Spark本地跑通Claude Code完全离线,隐私敏感开发者的新选择
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Optimizing GPT-OSS on NVIDIA DGX Spark: enabling local large model deployment. November 3, 2025. https://new.nat6.net/en/news/cmnxbjke50071sln0fe680fef/