如何获得对质量的“准”客观视角：OAIdeepresearch与Gemini的比较
为什么谷歌的DeepResearch比OpenAI的要好得多——我们现在是2025年5月，而DR OAI的“代理”甚至无法正确处理自己的模型……这只是一些不准确和不相关的信息，类似于下面的例子，这也是我认为Gemini会胜出的原因。我想？除非下面的质量差距是可以接受的……我认为这不可接受。我猜它唯一的主要工具就是网络搜索，但在进行研究时，它还没有弄清楚如何优先考虑最新的新闻……在进行深度研究时，考虑和优先处理最新的新闻和最新的博客对于许多用例来说是基本要求，除非我很愚蠢……比较这两个产品的研究片段……我们是否有资源测试最新版本在一系列广泛有用的案例中的表现，并进行并排比较，以查看哪个更好——或者有没有一些更定量的基准，能够捕捉研究输出的细微差别。
查看原文
Why is DeepResearch from Google this much better than from OpenAI – we are in 2025 May and DR OAI "Agent" cant even get its own model correct.. It’s just a bunch of inaccuracies and irrelevance that are similar to example below why i qualitatively think gemini is going to win out.. I guess?? Unless below gap in quality is acceptable.. I dont think it is. The only main tool i guess it has is web search and it hasn’t figured out how to prioritize recent news when it comes to doing research …. Factoring and favoring more recent news and latest blogs is kind of a basic requirement for deep research for a ton of use cases unless im stupid .. compare snippets from research from both products.. Do we have resources where the latest versions are tested across a range of broadly useful cases and do side by side comparison to see which one is better - or some more quntiative benchmark that can somehow capture nuance of research output.<p>https://i.imgur.com/9mPbURP.png<p>https://i.imgur.com/miwfF6H.png<p>https://i.imgur.com/8W9qwo8.png<p>https://i.imgur.com/kD02C4N.png