Kimi V4.1-1017 编程任务表现平平：想法多但常被丢弃
精选理由
有开发者实测 Kimi V4.1-1017 编程能力，说它想法很酷但老是半途丢弃，值得跑分前看看真实体验。
一位开发者实测 Kimi V4.1-1017 后认为，该模型在这类编程任务上相比此前没有明显提升，绘画能力也没有变化。测试中模型确实生成了更聪明的方案，也常常尝试更酷的做法，比如为天空场景自行构建 sky_lab.py，但最终大多被舍弃。作者的评价是潜力被大量浪费。
原文 · Teortaxes
I don't think V4.1-1017 improves much on these kinds of tasks. It didn't become a better painter, either. It is massively smarter and often tests cooler things, but mainly discards them. Here, it had some amazing ideas for the sky (built sky_lab.py). lots of wasted potential https://t.co/yWOiMiSeCM
