Expired50% confidenceEventExact time
A six-account comparative test used a third-party usage tracking tool to estimate an additional 1.4–1.6× discrepancy.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
9/7/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 9/21/2026(expired)
Sources
Related Entities
Related Claims
Unverified基准测试显示tiktoken比HuggingFace的tokenizers库快3-6倍61% similarUnverifiedThe lowest-purchase-price mechanism was tested on 5–6 numbers, with a success rate of approximately one-third61% similarUnverifiedGLM-5.3在六项评测中的成绩为：终端编程基准3.0得28.3、深度软件工程基准得66.9、智能体中场考试得28.5、自动化基准得48.2、带工具的人类终极考试得62.5、GDP价值评估AA2版得17.6960% similarUnverified研究表明当模型参数规模达到3B以上时，1.58-bit模型在多项基准测试中表现可接近甚至匹配同等规模的全精度模型58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/871260API
curl https://kongchang.com/api/v1/knowledge/claims/871260MCP
get_claim(id=871260)