Openai揭露了Chatgpt的“深入研究”功能，旨在攻读中国AI平台DeepSeek

随着与中国AI平台DeepSeek的竞争加剧，Openai为Chatgpt添加了一种新的“深度研究”模式。新功能允许用户

OpenAI has unveiled a new “deep research” mode for ChatGPT, allowing users to conduct multi-step research for complex tasks.

Openai推出了一种新的“深度研究”模式，用于gatgpt，使用户可以为复杂的任务进行多步研究。

The new feature, announced by OpenAI CEO Sam Altman on Friday, enables ChatGPT to perform complex research tasks on the web and generate a report.

OpenAI首席执行官Sam Altman周五宣布的新功能使Chatgpt能够在网络上执行复杂的研究任务并生成报告。

“today we launch deep research, our next agent. this is like a superpower; experts on demand!” Altman tweeted.

“今天，我们启动了深层研究，我们的下一个代理商。这就像超级大国；专家按需！”奥特曼发推文。

“it can go use the internet, do complex research and reasoning, and give you back a report. it is really good, and can do tasks that would take hours/days and cost hundreds of dollars.”

“它可以使用互联网，进行复杂的研究和推理，并给您报告。这真的很好，可以完成需要数小时/天的任务，并且花费数百美元。”

Deep Research is billed as OpenAI’s second AI agent, following the launch of Operator AI for browser-related tasks last month. It is powered by OpenAI’s latest o3 reasoning model, which is optimized for web browsing and data analysis.

在上个月推出了与浏览器相关的任务的操作员AI之后，深入研究被称为OpenAI的第二个AI代理。它由OpenAI的最新O3推理模型提供动力，该模型已针对Web浏览和数据分析进行了优化。

According to OpenAI, Deep Research can search, interpret, and analyze large amounts of text, images, and PDFs on the web to generate a comprehensive report, bringing its capabilities closer to those of a research analyst.

根据Openai的说法，深入的研究可以在网络上搜索，解释和分析大量文本，图像和PDF，以生成全面的报告，从而使其能力更接近研究分析师。

The company claims that deep research queries will take between 5 and 30 minutes to complete, and the chatbot will notify users when their research is finished.

该公司声称，深入的研究查询将需要5到30分钟的时间才能完成，聊天机器人将在完成研究完成后通知用户。

While Deep Research currently only supports text output, OpenAI plans to introduce embedded images, data visualizations, and other analytical output in the coming weeks.

尽管目前深入研究仅支持文本输出，但OpenAI计划在未来几周内引入嵌入式图像，数据可视化和其他分析输出。

In an X post, Altman provided a brief example of how Deep Research works.

在X帖子中，Altman提供了一个简短的示例，说明了深入研究的工作方式。

“people will post lots of great examples, but here is a fun one: i am in japan right now and looking for an old NSX. i spent hours searching unsuccessfully for the perfect one. i was about to give up and deep research just…found it,” he wrote.

“人们会发布很多很棒的例子，但这是一个有趣的例子：我现在在日本，正在寻找旧的NSX。我花了几个小时的时间搜索完美的搜索。我正准备放弃和深入研究……找到了。”他写道。

“it is very compute-intensive and slow, but it’s the first ai system that can do such a wide variety of complex, valuable tasks. going live in our pro tier now, with 100 queries per month. plus, team, and enterprise will come soon, and then free tier.”

“这是非常密集且缓慢的，但这是第一个可以执行如此多种复杂，有价值的任务的AI系统。现在在我们的职业层中生活，每月有100个查询。另外，团队和企业将很快到来，然后免费。”

However, OpenAI has also noted in its blog that deep research, like other AI tools from the company, has been known to “hallucinate.”

但是，Openai还在其博客中指出，与公司的其他AI工具一样，深入的研究是“幻觉”。

“Deep research may struggle to differentiate authoritative information from rumor, and currently exhibits weakness in confidence calibration, often failing to convey uncertainty accurately,” the blog post reads.

博客文章写道：“深入的研究可能难以将权威信息与谣言区分开来，目前在信心校准方面表现出弱点，通常无法准确传达不确定性。”