Lawyer with too many GPUs, part 5: DeepSeek-V4.1-Flash at near-NVFP4 quality on eight V100s from 2018. 524K context, 100+ tok/s decode, ~160 tok/s with two chats at once, and my legal harness now runs on it, 100% local

1. 

   消息推送

   Lawyer with too many GPUs, part 5: DeepSeek-V4.1-Flash at near-NVFP4 quality on eight V100s from 2018. 524K context, 100+ tok/s decode, ~160 tok/s with two chats at once, and my legal harness now runs on it, 100% local

   52分钟前

   Lawyer with too many GPUs, part 5: DeepSeek-V4.1-Flash at near-NVFP4 quality on eight V100s from 2018. 524K context, 100+ tok/s decode, ~160 tok/s with two chats at once, and my legal harness now runs on it, 100% local
   
   🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1xus81t/lawyer_with_too_many_gpus_part_5_deepseekv41flash/
   🏷️ 分组: Reddit｜AI·自托管·VPS
   🕒 时间: 2026-11-13 07:31:29

   Reddit

   

   From the LocalLLaMA community on Reddit: Lawyer with too many GPUs, part 5: DeepSeek-V4.1-Flash at near-NVFP4 quality on eight…

   Explore this post and more from the LocalLLaMA community

   52分钟前