嗯，用户让我帮忙总结一下这篇文章的内容，控制在一百个字以内。首先，我需要仔细阅读文章，理解其主要内容。文章主要讲的是微软正在将GPT 5.2集成到Copilot中，作为免费升级，同时与GPT 5.1共存。GPT 5.2被描述为OpenAI迄今为止最强的模型系列之一，能够更快完成实际任务，比如制作表格、编写代码、处理长文档等。 接下来，我注意到Copilot之前已经更新到了GPT 5.1的Smart模式，现在新增了GPT 5.2的“Smart Plus”模式。微软提到GPT 5.2在处理复杂任务方面表现更好，在知识工作和编程任务中的表现显著提升。具体来说，在知识工作方面，GPT 5.2击败或与行业专业人士持平的比例达到了70.9%，而之前的GPT 5只有38.8%。在编程方面，GPT 5.2在SWE-Bench Pro和Verified上的得分也比之前的版本高。 此外，文章还提到了一些基准测试结果，显示GPT 5.2在GPQA、AIME、CharXiv Reasoning和ARC-AGI等测试中的得分都很高。这些数据表明GPT 5.2在多个领域的能力都有显著提升。 现在我需要将这些信息浓缩到一百个字以内。首先要点出微软推出的是GPT 5.2集成到Copilot中，并且是免费升级。然后说明它比之前的版本更强，能够更快完成各种实际任务，并且在知识工作和编程方面表现突出。最后可以提到它在多个基准测试中表现出色。 所以总结起来就是：微软将GPT-5.2集成到Copilot中作为免费升级，该模型比前代更强大，在处理复杂任务、知识工作和编程方面表现出色，并且在多个基准测试中得分优异。 </think> 微软将GPT-5.2集成到Copilot中作为免费升级,该模型比前代更强大,可更快完成复杂任务,并在知识工作和编程方面表现出色,多个基准测试得分优异。 2025-12-29 12:30:28 Author: www.bleepingcomputer.com(查看原文) 阅读量:1 收藏
Microsoft is rolling out GPT 5.2 to Copilot on the web, Windows, and mobile as a free upgrade, and it'll coexist with the GPT 5.1 model.
GPT-5.2 is described as OpenAI’s strongest model series so far, and it allows you to finish real tasks faster, including things like building spreadsheets and presentations, writing and reviewing code, understanding long documents, using tools, and working with images.
Source: BleepingComputer
Last month, Copilot was updated with GPT 5.1 for the Smart mode. Now, Microsoft is adding GPT 5.2 to Copilot as a new "Smart Plus" mode.
GPT 5.2 on Copilot appears to be a 'Thinking' variant of GPT 5.2, as Microsoft clearly says it can handle complex tasks better.
On “knowledge work” tasks (GDPval, across 44 occupations), GPT-5.2 Thinking beats or ties industry professionals 70.9% of the time, compared to 38.8% for GPT-5.
That’s a huge jump, and it’s why OpenAI frames GPT-5.2 as “expert-level” for many well-specified office tasks (deliverables like decks, schedules, models, etc.).
For coding, GPT-5.2 Thinking posts 55.6% on SWE-Bench Pro and 80% on SWE-bench Verified, both higher than GPT-5.1 Thinking.
As per benchmarks, GPT 5.2 has GPQA Diamond 92.4%, AIME 2025 100%, CharXiv Reasoning (with Python) 88.7%, and higher scores on ARC-AGI (including a big jump on ARC-AGI-2).
Break down IAM silos like Bitpanda, KnowBe4, and PathAI
Broken IAM isn't just an IT problem - the impact ripples across your whole business.
This practical guide covers why traditional IAM practices fail to keep up with modern demands, examples of what "good" IAM looks like, and a simple checklist for building a scalable strategy.
如有侵权请联系:admin#unsafe.sh