<![CDATA[Easy_API的博客]]>https://blog.csdn.net/Easy_APIzh-cnhttps://blog.csdn.net/<![CDATA[总量涨了32倍，中位数没动]]>https://blog.csdn.net/Easy_API/article/details/169538427https://blog.csdn.net/Easy_API/article/details/169538427Easy_APIFri, 30 Oct 2026 09:56:57 +0800<![CDATA[2.8万亿参数，编程反超GPT-6 Astra：Kimi K3.1 Pro把开源模型推到了哪一步]]>https://blog.csdn.net/Easy_API/article/details/169455051https://blog.csdn.net/Easy_API/article/details/169455051Easy_APIThu, 29 Oct 2026 14:12:35 +0800<![CDATA[榜单前五只差5分，同一家的档位能差23分]]>https://blog.csdn.net/Easy_API/article/details/169446302https://blog.csdn.net/Easy_API/article/details/169446302Easy_APIThu, 29 Oct 2026 11:05:42 +0800<![CDATA[1GB的开源模型，把API月费省了]]>https://blog.csdn.net/Easy_API/article/details/169343850https://blog.csdn.net/Easy_API/article/details/169343850Easy_APIWed, 28 Oct 2026 10:25:30 +0800<![CDATA[同一个模型，六份账单：为什么关掉推理反而更贵？]]>https://blog.csdn.net/Easy_API/article/details/169257889https://blog.csdn.net/Easy_API/article/details/169257889Easy_APITue, 27 Oct 2026 13:54:58 +0800<![CDATA[同样的标价，单任务成本差了近三倍]]>https://blog.csdn.net/Easy_API/article/details/169243726https://blog.csdn.net/Easy_API/article/details/169243726Easy_APITue, 27 Oct 2026 10:09:52 +0800<![CDATA[同一套 API，两份价目表]]>https://blog.csdn.net/Easy_API/article/details/169164408https://blog.csdn.net/Easy_API/article/details/169164408Easy_APIMon, 26 Oct 2026 16:33:01 +0800<![CDATA[路由省下来的钱，有一部分是从质量里挪的]]>https://blog.csdn.net/Easy_API/article/details/169160760https://blog.csdn.net/Easy_API/article/details/169160760Easy_APIMon, 26 Oct 2026 14:49:33 +0800<![CDATA[大模型API的“安全评估”义务，正在被重新定义]]>https://blog.csdn.net/Easy_API/article/details/165585223https://blog.csdn.net/Easy_API/article/details/165585223Easy_APIWed, 16 Sep 2026 13:40:29 +0800<![CDATA[OpenRouter 9月首周编程榜出炉：国产模型首次包揽Top 3]]>https://blog.csdn.net/Easy_API/article/details/164860329https://blog.csdn.net/Easy_API/article/details/164860329Easy_APIThu, 10 Sep 2026 16:28:01 +0800