AnalysisAI ModelsOctober 21, 2026

LocalLLaMA users test DeepSeek-V4.1-Flash quants and benchmarks

Read original source →reddit.com

A €2k PC with an RTX 3090 24GB and 128GB DDR4 ran the 748B DeepSeek-V4.1-Flash-1017 at 7.2 tokens/s using the smallest 1-bit quant. Another user questions ISTA's GSQ-RCO 2-bit GGUFs, whose IQ2_XXS benchmarks match the base model with AIME25 at 100 for every version.

How this story unfolded

5 days · 0 reports · 3 community posts · from Oct 20

  1. Oct 20
  2. Oct 21
  3. Oct 25

More stories today

Open the live feed