Unverified50% confidenceBenchmarkExact time
DeepSeek-V4-Pro's 1.6T MoE requires the 1,440 GB memory pool of 8×B200 Blackwell under FP8.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/14/2026
First Seen
Valid until: 10/28/2026
Sources
Related Entities
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.75% similarVerifiedThe DeepSeek R1 7B version requires approximately 14 GB of storage at original FP16 precision, and only 4.7 GB after quantization.74% similarUnverifiedDeepSeek-V3 has a total of 685B parameters, and even after Q4 quantization its size still exceeds 400GB.74% similarUnverifiedAccording to community-cited information, DeepSeek-V4.1-Flash has 552B backbone parameters, 196B Engram parameters, and safetensors weight files totaling approximately 350 GB.72% 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.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1027765API
curl https://kongchang.com/api/v1/knowledge/claims/1027765MCP
get_claim(id=1027765)