Expired50% confidenceFactExact time
Running Kimi K3 requires approximately 700GB of VRAM, and the community estimates it needs an H200 cluster worth $3 million or multiple data-center-grade devices.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/22/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/20/2026(expired)
Sources
Related Claims
VerifiedKimi K3 supports an ultra-long context window of 1 million tokens.67% similarUnverifiedKimi K2在FP16精度下仅模型权重就需要数百GB显存，INT4量化可将显存需求压缩至约四分之一63% similarUnverifiedLlama 3 8B模型（80亿参数）运行约需7GB显存62% similarUnverifiedThe KV Cache of a 70B-parameter model with 128K context support can require hundreds of gigabytes of VRAM at full load, exceeding the 80GB VRAM capacity of a single H100.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/587703API
curl https://kongchang.com/api/v1/knowledge/claims/587703MCP
get_claim(id=587703)