Unverified50% confidenceSolutionExact time
/clear completely wipes the current session and starts from zero, with the lowest cost but loss of all context; /compact has the model compress the existing conversation into a summary that retains key information and then continues, still incurring one call's cost but far less than having the whole original text repeatedly participate in subsequent requests.
1
Sources
50%
Confidence
Long-term
Relevance
11/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe length of state maintained in a session affects cost: a long session carries prior context in each round, while a short session with cold-start input is smaller but may sacrifice continuity, and both of these cost levers are held on the client side.70% similarUnverifiedThe longer the conversation, the more expensive it gets toward the end, because every time a sentence is sent, the model reads all the preceding content again from the start, with the parts already read counted as cache hits (lower price) and the parts not yet read charged at the original price.70% similarUnverifiedOn short tasks, the three context management approaches of continuing the conversation, clear, and compact cost about 1.5, 3, and 4.5 cents respectively, and clear is only worthwhile when the old conversation is both large and already finished.69% similarUnverifiedAfter a long interruption (cache expiry), compaction is actually cheaper than simply continuing the conversation, because each subsequent turn then reads only about 15k tokens instead of the entire context.68% similarUnverifiedTransmitting only semantic deltas rather than full context can compress the token cost of multi-agent communication to a minimum.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1175569API
curl https://kongchang.com/api/v1/knowledge/claims/1175569MCP
get_claim(id=1175569)