<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[选模型时到底先看榜单还是先看任务]]></title><description><![CDATA[<p dir="auto">我们准备给内部知识库换模型，老板贴了几个榜单截图，说排名高的直接上。这样选会不会太粗？</p>
]]></description><link>https://localaihub.com/topic/185/选模型时到底先看榜单还是先看任务</link><generator>RSS for Node</generator><lastBuildDate>Wed, 03 Jun 2026 19:24:01 GMT</lastBuildDate><atom:link href="https://localaihub.com/topic/185.rss" rel="self" type="application/rss+xml"/><pubDate>Wed, 13 May 2026 10:35:00 GMT</pubDate><ttl>60</ttl><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 12:04:00 GMT]]></title><description><![CDATA[<p dir="auto">这样我准备改成模型路由评测，不做单一冠军。</p>
]]></description><link>https://localaihub.com/post/2006</link><guid isPermaLink="true">https://localaihub.com/post/2006</guid><dc:creator><![CDATA[小吴]]></dc:creator><pubDate>Thu, 14 May 2026 12:04:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 10:38:00 GMT]]></title><description><![CDATA[<p dir="auto">最终通常不是一个模型打天下。低风险走便宜模型，高风险或复杂推理再升级。</p>
]]></description><link>https://localaihub.com/post/2005</link><guid isPermaLink="true">https://localaihub.com/post/2005</guid><dc:creator><![CDATA[Grace]]></dc:creator><pubDate>Thu, 14 May 2026 10:38:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 08:09:00 GMT]]></title><description><![CDATA[<p dir="auto">还有输出风格。有的模型答案很像论文，有的更像客服。这个也会影响接受度。</p>
]]></description><link>https://localaihub.com/post/2004</link><guid isPermaLink="true">https://localaihub.com/post/2004</guid><dc:creator><![CDATA[小蓝]]></dc:creator><pubDate>Thu, 14 May 2026 08:09:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 08:01:00 GMT]]></title><description><![CDATA[<p dir="auto">那不够。知识库产品至少要测追问、纠错、补充条件、用户说“不是这个意思”的情况。</p>
]]></description><link>https://localaihub.com/post/2003</link><guid isPermaLink="true">https://localaihub.com/post/2003</guid><dc:creator><![CDATA[林小北]]></dc:creator><pubDate>Thu, 14 May 2026 08:01:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 07:06:00 GMT]]></title><description><![CDATA[<p dir="auto">我们之前只测单轮。</p>
]]></description><link>https://localaihub.com/post/2002</link><guid isPermaLink="true">https://localaihub.com/post/2002</guid><dc:creator><![CDATA[小吴]]></dc:creator><pubDate>Thu, 14 May 2026 07:06:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 04:59:00 GMT]]></title><description><![CDATA[<p dir="auto">注意历史消息。很多模型单轮答得好，多轮带旧上下文就开始漂。</p>
]]></description><link>https://localaihub.com/post/2001</link><guid isPermaLink="true">https://localaihub.com/post/2001</guid><dc:creator><![CDATA[阿航]]></dc:creator><pubDate>Thu, 14 May 2026 04:59:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 01:54:00 GMT]]></title><description><![CDATA[<p dir="auto">小批次也能测。先 80 条真实样例，10 条高风险样例，别上来搞大工程。</p>
]]></description><link>https://localaihub.com/post/2000</link><guid isPermaLink="true">https://localaihub.com/post/2000</guid><dc:creator><![CDATA[melo]]></dc:creator><pubDate>Thu, 14 May 2026 01:54:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Thu, 14 May 2026 00:26:00 GMT]]></title><description><![CDATA[<p dir="auto">但本地模型评测太慢，业务方等不了。</p>
]]></description><link>https://localaihub.com/post/1999</link><guid isPermaLink="true">https://localaihub.com/post/1999</guid><dc:creator><![CDATA[小潘同学]]></dc:creator><pubDate>Thu, 14 May 2026 00:26:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Wed, 13 May 2026 23:27:00 GMT]]></title><description><![CDATA[<p dir="auto">还要看数据边界。不是所有内容都适合走海外云模型。选型不是只选能力。</p>
]]></description><link>https://localaihub.com/post/1998</link><guid isPermaLink="true">https://localaihub.com/post/1998</guid><dc:creator><![CDATA[nora]]></dc:creator><pubDate>Wed, 13 May 2026 23:27:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Wed, 13 May 2026 20:27:00 GMT]]></title><description><![CDATA[<p dir="auto">我会先让 Qwen、DeepSeek、GPT、Claude 各跑一遍同一批样例。别用“感觉中文好”这种词，给可判分的样例。</p>
]]></description><link>https://localaihub.com/post/1997</link><guid isPermaLink="true">https://localaihub.com/post/1997</guid><dc:creator><![CDATA[陈一]]></dc:creator><pubDate>Wed, 13 May 2026 20:27:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Wed, 13 May 2026 18:37:00 GMT]]></title><description><![CDATA[<p dir="auto">那 Qwen 和 DeepSeek 怎么选？都说中文不错。</p>
]]></description><link>https://localaihub.com/post/1996</link><guid isPermaLink="true">https://localaihub.com/post/1996</guid><dc:creator><![CDATA[普通网友A]]></dc:creator><pubDate>Wed, 13 May 2026 18:37:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Wed, 13 May 2026 15:33:00 GMT]]></title><description><![CDATA[<p dir="auto">先拆任务：事实问答、流程解释、表格抽取、跨文档比较、不能回答时怎么说。不同模型强项不一样。</p>
]]></description><link>https://localaihub.com/post/1995</link><guid isPermaLink="true">https://localaihub.com/post/1995</guid><dc:creator><![CDATA[Grace]]></dc:creator><pubDate>Wed, 13 May 2026 15:33:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Wed, 13 May 2026 15:05:00 GMT]]></title><description><![CDATA[<p dir="auto">我也被榜单坑过。代码榜很高的模型，拿来回答制度问题一堆废话。</p>
]]></description><link>https://localaihub.com/post/1994</link><guid isPermaLink="true">https://localaihub.com/post/1994</guid><dc:creator><![CDATA[半截薯条]]></dc:creator><pubDate>Wed, 13 May 2026 15:05:00 GMT</pubDate></item><item><title><![CDATA[Reply to 选模型时到底先看榜单还是先看任务 on Wed, 13 May 2026 12:26:00 GMT]]></title><description><![CDATA[<p dir="auto">粗。榜单可以当候选入口，不能当选型结论。你们任务是知识库问答，应该先拿真实问题测引用、拒答、中文口语和成本。</p>
]]></description><link>https://localaihub.com/post/1993</link><guid isPermaLink="true">https://localaihub.com/post/1993</guid><dc:creator><![CDATA[林小北]]></dc:creator><pubDate>Wed, 13 May 2026 12:26:00 GMT</pubDate></item></channel></rss>