Unverified50% confidenceBenchmarkExact time
On a dataset containing 893 BPMN XML files, tok models require 2.2 times fewer tokens than the smallest of the XML files and 7.7 times fewer than the original exported form of the files.
1
Sources
50%
Confidence
Long-term
Relevance
11/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn a 4000-token database dump, perhaps only 120 tokens contain useful business facts, and the remaining 3880 tokens serve as computational noise that dilutes attention.64% similarUnverifiedFor this architecture, the cached bytes per token is not a constant; the equivalent bytes/token of the window layers alone is about 688.0 at a context of 8,192, about 43.0 at 131,072, and about 5.4 at 1,048,576.60% similarUnverifiedThe study validates on three datasets—TinyStories, WikiText-103, and CNN/DailyMail—scaling each corpus from 512 chunks to 2,048 chunks, with each chunk consisting of 64 tokens.60% similarUnverifiedAnthropic要求缓存的最小单位是1024个Token，OpenAI以128个Token为粒度进行缓存匹配60% similarUnverifiedA log study measured the tokens each harness sends before reading the user's prompt: OpenCode approximately 7,000 tokens, CloudCode approximately 33,000 tokens, and Pi approximately 1,300 tokens.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1142418API
curl https://kongchang.com/api/v1/knowledge/claims/1142418MCP
get_claim(id=1142418)