Unverified50% confidenceBenchmarkExact time
In a 40-turn conversation test using claude-sonnet-5 (input $0.3835/1M, output $1.5339/1M), the full-history strategy cost $0.001599 per message, a sliding window of the most recent 6 turns cost $0.000413 (about −74%), and summary plus the most recent 4 turns cost $0.000388 (about −76%).
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/15/2026
First Seen
Valid until: 10/29/2026
Sources
Related Entities
Related Claims
UnverifiedIn Phase 1, Sonnet 5.5 took 4 hours and 50 minutes, made 1,670 calls, and incurred an equivalent bill of $164.79, of which cache reads cost $114.75% similarUnverifiedMainstream model inference costs are priced per million tokens: GPT-4o is approximately $5–15, Claude 3.5 Sonnet is approximately $3–15, and the total token count for multi-step agent tasks is often 10–50 times that of a single conversation.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1037060API
curl https://kongchang.com/api/v1/knowledge/claims/1037060MCP
get_claim(id=1037060)