Unverified60% confidenceBenchmarkExact time
After compression, the DeepSeek V4 Flash model size was reduced from nearly 500 GB to 80.8 GB.
2
Sources
60%
Confidence
Long-term
Relevance
7/25/2026
First Seen
Sources
Redis之父如何将2840亿参数DeepSeek塞进普通笔记本
Related Claims
UnverifiedOn MI355X, by removing unnecessary expert padding, the FP4 MoE model weights of DeepSeek-V4 were reduced from 159.07 GB to 112.36 GB.76% similarUnverifiedModels using MLA such as DeepSeek-V4.1-Flash compress the KV state to about 2.4 GB per 32k stream (FP8 KV), while wider architectures consume 3.0 GB–3.8 GB per 32k stream.75% similarVerifiedDeepSeek-V4.1-Flash's KV cache is compressed to 890 bytes per token, only one-quarter that of V4 Flash.74% similarUnverifiedThe UD-Q8_K_XL quantized version of DeepSeek V4 Flash has a file size of 162 GB, uses MXFP4+BF16 mixed precision, and achieves completely lossless quality.74% similarUnverifiedDeepSeek-V3 has a total of 685B parameters, and even after Q4 quantization its size still exceeds 400GB.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/613177API
curl https://kongchang.com/api/v1/knowledge/claims/613177MCP
get_claim(id=613177)