Unverified50% confidenceFactExact time
Each time the advisor model is invoked, it re-reads the entire session history from scratch, and the advisor's own reads are not cached, while the main model's cache remains intact.
1
Sources
50%
Confidence
Long-term
Relevance
11/16/2026
First Seen
Sources
Claude Code Advisor Mode: Two-Model Collaboration for Cost-Effective Coding
youtubeThirdHand AI English11/15/2026
Related Entities
Related Claims
UnverifiedWhen the advisor is invoked, it re-reads the full session as input, and this reading does not enter prompt caching, so each summon is billed at full price for input tokens; the main model's cache is unaffected, and the advisor's returned suggestions are cached locally in subsequent steps.79% similarUnverifiedWriting project rules into an external file and re-reading it in each conversation can substitute for the memory that AI lacks.71% similarUnverifiedThe advisor model itself does not write code or take over the keyboard; it only outputs strategic judgments or review opinions when invoked, and actual code execution is still performed by the main model.70% similarUnverifiedSome users suggested using the quota before a reset to have the AI summarize chat records, generate knowledge base summaries, or export important conversations, so that even if the quota resets, the core information has already been saved.69% similarUnverifiedA chat application can "remember" historical conversations because the backend database stores every message and sends the entire history together with the new message to the model on each new request; the model itself does not retain any state across calls.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1175527API
curl https://kongchang.com/api/v1/knowledge/claims/1175527MCP
get_claim(id=1175527)