Unverified50% confidenceBenchmarkExact time
A lean CLAUDE.md configuration sends 16,500 tokens, while a bloated version (356-line CLAUDE.md, 20 skills, 3 MCP servers) sends 21,500 tokens per message.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
Related Entities
Related Claims
UnverifiedUsing NVIDIA B300 chips in a 1-megawatt data center can achieve a total throughput of approximately 450,000 tokens/s, which at 340 tokens/s per user amounts to approximately 1,324 concurrent active users.68% similarUnverifiedResNet-50 has approximately 25 million parameters, while GPT-3 reaches 175 billion parameters.63% similarUnverifiedWhisper-Medium 拥有约 3 亿参数的编码器部分63% similarUnverifiedRecommended VRAM tiers for local deployment: 4–6 GB VRAM for ~4-billion-parameter models, 8 GB for ~9-billion-parameter models, 16 GB for ~14-billion-parameter models, and 24–32 GB for ~27-billion-parameter models.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/996115API
curl https://kongchang.com/api/v1/knowledge/claims/996115MCP
get_claim(id=996115)