{"version":"1.0","type":"card","id":"decee706-93c3-479d-a05b-95f260670c8e","url":"https://stacklist.com/card/decee706-93c3-479d-a05b-95f260670c8e","title":"Qwen3.8-Flash-Next Performance Improvements","source_url":"https://x.com/miaai_lab/status/2096601527456592223?s=12","note":"This page discusses the latest enhancements to Qwen3.8-Flash-Next running on a single NVIDIA DGX Spark. Key improvements include a 24.7% increase in prefill speed at 8k and a fix for a bug that previously reduced decode performance by up to 24%.","image":{"url":"https://ucarecdn.com/aaa76270-7734-475e-bb48-9a7ae8d97edc/","alt":"Qwen3.8-Flash-Next Performance Improvements","width":1035,"height":525},"stack":{"id":"cc483fc9-ce97-46ea-bc49-7a300af7cd39","title":"GitHub repos to check out","url":"https://stacklist.com/c/technology/stack/cc483fc9-ce97-46ea-bc49-7a300af7cd39"},"created_at":"2026-09-07T16:44:50.856Z","updated_at":null,"aco":null,"_links":{"self":"/api/public/card/decee706-93c3-479d-a05b-95f260670c8e.json","html":"https://stacklist.com/card/decee706-93c3-479d-a05b-95f260670c8e","md":"/api/public/card/decee706-93c3-479d-a05b-95f260670c8e.md","stack_json":"/api/public/stack/cc483fc9-ce97-46ea-bc49-7a300af7cd39.json"}}