Unverified50% confidenceBenchmarkExact time
The PagedAttention paper uses OPT-13B in FP16 as an example, where each token requires approximately 800 KB, and 2048 tokens require approximately 1.6 GB of KV Cache.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Verified以LLaMA-3 70B（80层、64头、每头128维）为例，单个token的KV Cache约占2.6MB（FP16精度），32K上下文的KV Cache总量约达83GB80% similarVerifiedWhen using FP16 precision, the KV Cache for LLaMA-2-7B occupies approximately 0.5 MB per token.73% similarUnverifiedAccording to community-cited information, DeepSeek-V4.1-Flash has 552B backbone parameters, 196B Engram parameters, and safetensors weight files totaling approximately 350 GB.71% similarVerifiedA GPT-4-class model with 175 billion parameters requires approximately 175 GB of VRAM even when stored at fp8 precision.71% similarVerifiedA 70B-parameter FP16 model requires approximately 140 GB of video memory, approximately 70 GB after INT8 quantization, and approximately 35–40 GB after INT4 quantization.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/908131API
curl https://kongchang.com/api/v1/knowledge/claims/908131MCP
get_claim(id=908131)