Unverified50% confidenceOpinionExact time
Opus 5.5在review作者的视频生成框架时指出整个流程写了1.4万行Python和JS，而真正决定画面的源码只有651行，其余大量是无用的测试
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/29/2026
First Seen
Valid until: 12/28/2026
Sources
Related Entities
Related Claims
Unverified最终结果为6/6数值测试通过，0个依赖漏洞，转换1,775个模块，生产JavaScript约213.6 kB，Gzip后约68.5 kB70% similarUnverified在Aider Polyglot编码测试中，DeepSeek全精度得71.6%，Unsloth 4-bit动态版得69.7%，2-bit版得65.8%，而同为2-bit同文件大小的竞品量化方案只得56.6%64% similarUnverifiedOpus 5在AI Omniscience测试拿了31分，避免瞎编的能力优于Opus 4.864% similarUnverifiedOpus 4.7每次解题的中位输出token数为60,000，而GPT 5.5仅需16,00062% similarUnverified微调后的406M模型在测试集上取得36.33 ROUGE-1，作者自建的1.2B系统为35.41 ROUGE-1，两者在统计意义上无法被区分62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/958885API
curl https://kongchang.com/api/v1/knowledge/claims/958885MCP
get_claim(id=958885)