Unverified50% confidenceBenchmarkExact time
在512-4096上下文长度测试中，将草稿长度从4字符拉满至16字符，整体推理延迟仅增加0.2%-1.3%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/17/2026
First Seen
Valid until: 10/15/2026
Sources
Related Claims
Unverified千问3.8 27B的代码方向进步明显大于文本创作，剧本拆解能力相比3.6有提升但幅度有限69% similarUnverified标称0.1g精度但5kg大量程的秤要警惕，量程越大分度值越大时误差越明显，真正0.1g稳定读数的一般量程不超过3kg64% similarUnverified草稿长度为15字符时，DSpark在数学、代码、对话场景的效率提升分别达到30%、26%、22%64% similarUnverified在中文字数控制测试中，要求恰好写40字，Claude Sonnet 4.5精确完成40字，GPT 5.1写了37字61% similarUnverified早期在Opus模型上测试精确字符串识别率为15分之0，而Fable 5发布后同样测试变成15分之1361% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/541581API
curl https://kongchang.com/api/v1/knowledge/claims/541581MCP
get_claim(id=541581)