Primary source掘金本周最热
View sourceEvent ·Multi-source
Agent场景下谁才是好用的Flash模型?Step 3.7 Flash实测
文章对比了Step 3.7 Flash、DeepSeek V4 Flash、Gemini 3.5 Flash在Agent场景下的代码生成效率、响应速度和工具调用稳定性。测试采用Claude Code工具,第一个案例从零搭建开发者日志站,Step 3.7 Flash一次生成,输出25.7k tokens,成本¥1.22,耗时2m30s;DeepSeek V4 Flash成本¥0.72,输出14k tokens。第二个案例搭建GitHub项目雷达,Step 3.7 Flash无错误完成,Gemini 3.5 Flash有2次自动修复报错。在视觉效果和稳定性上Step 3.7 Flash表现更优,但DeepSeek V4 Flash成本更低。
Current read
最近这几个Flash模型我帮你试了,Step 3.7 Flash写博客页和抓GitHub项目都很稳,页面好看,一次跑通,成本也就一两块钱,值得试试。
3 sources51° AI heatLast updated 2026/6/29 15:11:43
Evidence
3 sourcesConflict check
Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.