antirez releases DeepSeek V4 Flash GGUF quantizations, from 2-bit to 4-bit, for local deployment

antirez · x · 2026-08-01

Redis creator antirez has released GGUF quantizations of DeepSeek V4 Flash on Hugging Face, including IQ2XXS, Q2K, Q4K, mixed-precision, and MTP versions, totaling 2.79 TB. The files have passed smoke tests but not full testing, with antirez planning to test them tomorrow.

Original post →

More from Models

Models channel →