Unverified50% confidenceBenchmarkExact time
In Perplexity's RAG scenarios, a single user query may trigger 20–50 model calls.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/9/2026
First Seen
Valid until: 12/8/2026
Sources
Related Entities
Related Claims
Unverified超低输入价格对RAG场景有利，因为典型RAG查询可能注入5000-20000个token的检索结果而输出仅有200-500个token68% similarUnverifiedIndustry estimates suggest that implicit tokens (system prompts, RAG retrieval context, function descriptions, and conversation history) typically account for 60%–80% of total consumption per request.66% similarUnverified推理模型通过思维链或内部搜索进行多步推演，token消耗量可能是普通模型的5到20倍62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/884076API
curl https://kongchang.com/api/v1/knowledge/claims/884076MCP
get_claim(id=884076)