跳到正文
Hacker News:AI 熱帖·· 19 天前AI 評分31

開放模型能以二十分之一成本完成九成智能體工作

Open models can do 90 percent of agentic work at a twentieth of the cost

AI 導讀

開放模型可完成九成智能體工作,成本僅為原成本的二十分之一。評論者指出,開放權重模型可能較冗長、效率較低;自行託管時則是另一回事。

正文

Image 1Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Open models can do 90 percent of agentic work at a twentieth of the cost (theregister.com) 21 points by Betelbuddy8 hours ago | hide | past | favorite | 4 comments help Image 2elijahglass2 hours ago | next[–] Open weight models are extremely capable. The stats the article relies on are not the most accurate since open weight models are known for higher verbosity and lack of efficiency. Although if you're self hosting open weights that's a different story all together. reply Image 3Northernator116 hours ago | prev[–] New deep seek 4.1 has done the majority of coding tasks ive thrown at it no problem reply Image 4phillc735 hours ago | parent[–] Deepseek 4.1 Flash 1017 has been pretty good for coding, I agree. It doesn't seem to do documentation as well, things like requirements, specifications, implementation plans. Here I still use GLM5.3 for a better outcome. I also use both models as adversarial batteries to check the work of each other. Having said that, 1017 has been great in saving money. It's available generally much cheaper than GLM5.3, and I think it delivers >90% of the capability. reply Image 5Northernator1154 minutes ago | root | parent[–] Will have to try this review stratergy with GLM, Thanks for sharing reply Image 6 Consider applying for YC's Winter 2027 batch! Applications are open till November 2. Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact Search:

來源:Hacker News:AI 熱帖 · news.ycombinator.com