Unverified50% confidenceFactExact time
一次典型的RAG调用可能检索5-10个文档块，每块500-1000个Token，输入Token总量可达5000-10000个，而生成回答通常只有200-500个Token
1
Sources
50%
Confidence
Long-term
Relevance
8/28/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedDocument chunks retrieved in each RAG call typically range from 500 to 2,000 tokens, and can reach several thousand tokens when multiple retrieval paths are used.68% similarUnverifiedThe batch-scoring step processes approximately 10 abstracts per batch, totaling about 18,000 tokens.68% similarUnverifiedThe description text for 60 Skills/MCPs totals approximately 12,000 tokens (an average of 200 tokens each).67% similarUnverified在一个典型项目查询中发送了45,000 tokens的上下文，但真正有用的只有约5,000 tokens66% similarUnverifiedA single code review or generation request typically consumes approximately 500–2,000 tokens.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/810514API
curl https://kongchang.com/api/v1/knowledge/claims/810514MCP
get_claim(id=810514)