Opus 5.5 与 GPT-6 Sol 和 Luna

这周前半段，我们看到了不止一个、不止两个、不止三个，而是四个重大模型发布，包括本周二的两个，还有OpenAI的两个和Anthropic的一个。

The first half of this week has seen not one, not two, not three, but four major model releases, including two on Tuesday this week, a pair from OpenAI and one from Anthropic.

对于OpenAI来说，GPT-6 Sol和Luna继续他们在智能堆栈各个层面构建成本高效模型的追求；而对于Anthropic，早期迹象表明Opus 5.5是回归荣耀，或者至少是自Opus 4.6以来公司未曾见过的早期采用者的赞誉。

For OpenAI, GPT-6 Sol and Luna continue their quest to build cost-efficient models at every level of the intelligence stack, and for Anthropic, early indications suggest that Opus 5.5 is a return to glory, or at least early adopter acclaim that the company has not seen since Opus 4.6.

《AI每日简报》是一个每日播客和视频节目，涵盖AI领域最重要的新闻和讨论。

The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.

好了，朋友们，在我们深入之前，先快速说几个公告。

All right, friends, quick announcements before we dive in.

首先，感谢今天的赞助商：KPMG、Blitzy、Section和Hyperagent。

First of all, thank you to today's sponsors, KPMG, Blitzy, Section, and Hyperagent.

要收听无广告版本的节目，请访问patreon.com/aidailybrief，或者您也可以在Apple Podcasts上订阅。

To get an ad-free version of the show, go to patreon.com/aidailybrief, or you can subscribe on Apple Podcasts.

想了解更多关于赞助节目的信息，请发送邮件至sponsors@aidailybrief.ai。

And to learn more about sponsoring the show, send us a note at sponsors@aidailybrief.ai.

现在，像往常一样，当有重大新模型发布时，这期节目将只包含主要内容。

Now, as is normal when it comes to big new model releases, this will be a main-only episode.

说实话，就算这样，我们想把所有内容都讲完也挺难的，所以咱们直接开始吧。

And honestly, we're going to have a hard time getting it all in even with that, so let's dive in.

欢迎回到AI每日简报。

Welcome back to the AI Daily Brief.

不可否认，AI每日简报上最精彩的日子，当然也是最有趣、最激动人心、最充满活力的那种“哎呀，我迫不及待想录完这期节目，因为接下来我就可以去干别的事了”的节目，都是我们拿到新模型的时候。

It's fairly undeniable that the best days on the AI Daily Brief, certainly the most fun, exciting, dynamic, "Oh, boy, I can't wait to be done with this episode because now I get to go do things" types of episodes are ones where we get new models.

而昨天，我们遇到了极其罕见的待遇，我几乎不记得以前发生过这种事，那就是当下最重要的两个实验室，OpenAI和Anthropic，在同一天发布了新模型。

And yesterday, we got the absolute rarest of treats, a thing which I can't frankly remember ever happening before, which is the two most important labs of the moment, OpenAI and Anthropic, both releasing new models on the same day.

当然，以前也有过模型发布时间很接近的情况。

Certainly, we've had models released close together before.

实际上，我们通常看到的模式是像SpaceX AI昨天那样，提前发布他们的模型，以免被他们知道会获得更多关注的模型所淹没。

In fact, usually the pattern that we see is what SpaceX AI did yesterday, releasing their model in advance so as not to be drowned out by models that they knew would get more attention than them.

同日发布的挑战在于，它不可避免地让人们不再问这个模型有什么价值、它在我日常使用中会扮演什么角色，而是问这两个哪个更好，以及这说明了哪个实验室处于领先地位？

The challenge of the same-day release is that it inevitably gets people to ask not what's valuable about this particular model and where it's going to fit in my rotation, but instead, which of these is better, and what does it say about which lab is in the lead?

所以今天，我们要来梳理一下发布了什么、基准测试的情况、初步反应、一些特定用例的例子、对竞争格局的影响、这对整个前沿推进速度意味着什么，以及根据社区的估计，谁赢得了这一天。

So today, we are going to go through what was released, where things stand on the benchmarks, the first reactions, examples around some particular use cases, the impact on the competitive landscape, what it says about the whole pacing the frontier thing, and, in the community's estimation, who won the day.

我们拿到的第一个模型是 Claude Opus 5.5，说实话，已经很久没有 Opus 系列的模型成为焦点了。

The first model we got was Claude Opus 5.5, and frankly, it's been some time since an Opus class model was the big show.

当然，最主要的原因就是 Mythos 和 Fable 的推出。

Now, the biggest reason for that is, of course, the introduction of Mythos and then Fable.

但在那之前，大家对 Opus 4.6 的评价就非常高，所以后来的 4.7 和 4.8 对很多人来说，如果不是倒退的话，充其量也就是小幅升级，甚至在某些方面还不如从前。

But even before that, people had so much love for Opus 4.6 that 4.7 and 4.8 were for many, if not a regression, certainly very incremental at best, and not even incrementally ahead in certain cases.

事实上，到了 Opus 5 的时候，大家是真的完全不喜欢这个版本。

And in fact, when it came to Opus 5, people just genuinely did not like the thing at all.

接着说回大家最关心的部分，AI内容创作者Peter Yang做了一个特别形象的比喻。他用一匹马的插画来代表Claude的不同版本：一开始是精美完整的画作，对应Opus 4.6；然后画风变得粗糙、简单甚至有点童趣，对应Opus 4.7和4.8；到了Opus 5几乎成了乱涂乱画；最后又回到了Opus 5.5，画面再次变得精美细致，马的正面轮廓完美呈现。

Now, jumping ahead to where this conversation is landing, the perfect encapsulation comes from AI content creator Peter Yang, who used a meme of an illustration of a horse, beautiful and complete at the beginning, representing Opus 4.6, of course, to a much scratchier, more simplistic and childlike line drawing for Opus 4.7 and Opus 4.8, culminating in near scribbles for Opus 5, to finally once again, the beautiful front side of a horse in perfect illustrated detail for Opus 5.5.

也就是说，大家真的很喜欢这个新模型。

The people, in other words, are really liking this model.

不过Anthropic是怎么宣传它的呢？

But how did Anthropic pitch it?

在他们的公告里提到，Opus 5.5在大多数任务上的表现能达到Claude Fable 5.1的水平，但运行成本降低了40%。这不仅仅是比Fable 5.1便宜，甚至比之前的Opus 5还要便宜。

In their announcement post, they said that Opus 5.5 performs at the level of Claude Fable 5.1 for most tasks but costs 40% less to run, not even just than Fable 5.1, but than Opus 5.

在他们的公告帖子里，他们指出这相比Opus 5是一个重大进步，但坦白说，至少在基准测试方面，它也比Fable 5.1更胜一筹。

In their announcement thread, they point out that it is a major step up from Opus 5, but also frankly, at least when it comes to the benchmarks, it's also a step up from Fable 5.1.

在Terminal Bench 4.0上，模型从Fable 5.1的55.8%跃升到Opus 5.5的66.4%。

On Terminal Bench 4.0, the model jumped from Fable 5.1's 55.8% to Opus 5.5's 66.4%.

Cursorbench、Frontier Code v1.1和Humanity's Last Exam也都出现了提升，不仅超过了Opus 5，也超过了Fable 5.1。

Cursorbench, Frontier Code v1.1, and Humanity's Last Exam also all saw jumps, not only from Opus 5, but from Fable 5.1 as well.

事实上，没有一个基准测试是Opus 5.5落后于其他Anthropic模型（Fable 5.1和Opus 5）的，只有两个基准测试——Automation Bench（衡量业务流程）和Terminal Bench Science（衡量代理式科学研究）——Opus 5.5没有领先于GPT-6 Astra。

In fact, there wasn't a single benchmark that Opus 5.5 wasn't ahead of the other Anthropic models, Fable 5.1 and Opus 5, and only two, Automation Bench, which measures business workflows, and Terminal Bench Science, which measures agentic scientific research, where Opus 5.5 wasn't ahead of GPT-6 Astra as well.

重要的是，这不仅是性能上的提升，也是价格上的优势。

Importantly, this is not just a performance gain, but also a pricing gain as well.

Claude账号写道，因为Opus 5.5运行所需的计算资源比Opus 5少，所以它的定价也随之降低。

The Claude account wrote that because Opus 5.5 requires less compute to serve than Opus 5, its pricing is consequently down.

相比Opus 5每百万输入令牌5美元和每百万输出令牌25美元，Opus 5.5是每百万输入4美元和每百万输出20美元。

Compared to Opus 5's $5 per million input tokens and $25 per million output, Opus 5.5 is $4 per million input and $20 per million output.

然而，由于额外的效率提升，他们说实际成本节省将达到约40%，而不仅仅是价格上反映出的20%。

However, because of additional efficiencies, they said that the cost gains will actually be about 40% as opposed to just the 20% reflected in the cost.

虽然它主打的不是速度，但他们也提到，因为很多默认设置下的表现都比其他模型在高设置下还要好，所以 Opus 5.5 的输出速度比那些模型快得多，平均比 Opus 5 快了 30%。

Although it's not pitched as a model focused on speed, they do note that because many of its default effort settings deliver better results than other models running at higher settings, that Opus 5.5 generates output much faster than those other models, including an average of 30% faster than Opus 5.

更棒的是，他们提高了所有 Pro、Max 和 Team 套餐的五小时使用时长限制，还给订阅用户增加了速率限制重置功能，大家随时都能用。

As a cherry on top, they increased five-hour usage limits on all of their Pro, Max, and Team plans, and threw in the rest of the subscription users a rate limit reset, which can be used whenever people need.

在安全方面，他们指出这是他们在发布《把控前沿》报告后推出的首款模型，Opus 5.5 在发布前经过了包括 Frontier Design 和 Meter 在内的外部评估机构的测试。总的来说，他们认为 Opus 5.5 基本上是他们目前发布的对齐效果最好的模型。

On the safety front, they note that this is their first model released in the wake of the Pacing the Frontier note, and that Opus 5.5 was tested before release by external evaluators, including Frontier Design and Meter, and overall, they argued that Opus 5.5 is basically the best aligned model they've released yet.

Anthropic 负责对齐工作的 Sam Bowman 写道："我们认为 Opus 5.5 比其前身更安全得多，因此发布它极有可能降低与不对齐相关的风险。"

Sam Bowman, who works on alignment at Anthropic, wrote, "We think Opus 5.5 is sufficiently safer than its predecessors that releasing it, more likely than not, reduces risks related to misalignment."

光凭 Opus 5.5 就足以撑起一整期节目了，但 OpenAI 显然不想让 Claude 独占风头。周二美东时间下午早些时候，他们自己也甩出了一套新模型：GPT-6 Sol 和 GPT-6 Luna。

Now, Opus 5.5 alone would be more than enough for a complete episode, but OpenAI was determined not to let Claude have all the fun, and in the early afternoon Eastern Time on Tuesday, they dropped their own set of models, GPT-6 Sol and GPT-6 Luna.

如果说 Opus 5.5 的发布还多少带点对成本和效率的关注，那这次 Sol 和 Luna 的发布简直就是把“成本”和“效率”这两个词刻在了脑门上。

And if the Opus 5.5 announcement was a little bit focused on cost and efficiency, the 6 Sol and 6 Luna announcement was all about cost and efficiency.

他们在 X 平台上的公告里写道，GPT-6 Sol 和 Luna 建立在 GPT-6 Astra 的技术突破之上，将其核心优势转化为了更快、更实惠的模型，以支持大规模应用。

Their announcement post on X reads GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, bringing much of its strengths into faster and more affordable models to support work at scale.

我们还优化了缓存和推理的效率，并且把这些省下来的成本直接让利给你们。

We've also made caching and inference more efficient, and we're passing the savings directly to you.

Sol 和 Luna 的 API 价格比 GPT-5.6 的促销价低了 50%。

50% lower API prices for Sol and Luna compared with GPT-5.6 promotional pricing.

也就是说，他们不只是说这个价格比 Astra 便宜，而是在上一代的基础上就已经在降价了。

In other words, they're not just talking about this pricing being lower than Astra, they're cutting costs from the previous generation even before that.

有意思的是，他们明显也在把成本降低说成是解锁新用法的方式。

And what's interesting is that they're clearly presenting cost as enabling new ways to use the tool as well.

他们说的意思是，更高的使用限制和更低的成本给了你更多的灵活性和迭代空间。

As they put it, higher usage limits and lower cost give you more flexibility and room to iterate.

换句话说，他们认为人们利用这些成本节省的方式之一，就是更频繁地迭代使用模型。

In other words, one of the ways that they think that people should cash in on those cost savings is to be more iterative in how they use the model.

回顾几代模型，Codex 和 GPT 的使用方式，与 Claude 和 Fable 的使用方式之间，一直有一个微妙的区别：人们倾向于用 Fable 和 Claude 来处理那些不需要太多干预、可以长时间运行的任务，而在 Codex 里用 GPT 模型来处理那些需要更多交互和迭代的事情。而这次新的 Sol 和 Luna 模型似乎也延续了这个思路。

Going back a couple of generations of models at this point, one of the subtle divides between Codex and GPT usage, and Claude and Fable usage on the other hand, has been an impulse to use Fable and Claude for more hands-off long-running tasks, and the GPT models inside Codex for things where you need more interaction and iteration, and these new Sol and Luna models seem to build on that as well.

和 Opus 5.5 的发布一样，OpenAI 也提到了在安全对齐方面的进展。

Like the Opus 5.5 announcement, OpenAI also points to advancements in alignment.

值得注意的是，当你去看 OpenAI 页面上的基准测试时，Anthropic 还在用传统的图表，显示基准名称和一系列模型上的百分比分数，而 OpenAI 已经完全抛弃了那种方式，只在图表上展示基准测试，其中 y 轴是性能，x 轴是成本。

And what's notable when you get into the benchmarks on the OpenAI page is that where Anthropic is still using the traditional charts, which show a benchmark name and then a set of percentage scores across a set of models, OpenAI has entirely abandoned that, only showing benchmarks on graphs that map performance on the y-axis versus cost on the x-axis.

山姆·奥特曼发推说：“GPT-6 Sol 和 Luna 在智能、对齐、工作产出、编程、计算机使用等方面，相比 5/6 系列的前代产品都有大幅提升。”

Sam Altman tweeted, "GPT-6 Sol and Luna are big improvements on intelligence, alignment, work output, coding, computer use, and more over their 5/6 family predecessors.

“而且它们的每 token 价格只有一半，每任务成本甚至更低。”

They are also half the price per token and even less per task."

然后他又接着强调，这种思维方式对现在的 OpenAI 有多重要，尤其是按任务定价这个指标才是真正重要的。山姆说：“我觉得市场上没有任何东西能跟我们竞争。”

He then went and followed up and really reinforced just how important this way of thinking is to OpenAI now, especially compared by per task pricing, which is the metric that should matter, Sam says, "I don't think there is anything competitive anywhere in the market.

“我们希望人们能大量使用 AI。”

We want people to be able to use tons of AI.

“这对探索我们面前这个新文艺复兴至关重要。”

It's important to being able to explore this new renaissance in front of us."

为了让大家对“大量”有个概念，他们在发布公告里指出，按 API 价格计算，OpenAI 的中位数研究员每天消耗 600 美元的 token，而第 90 百分位的研究员每天消耗 7000 美元的 token。

Putting some magnitude around what they mean by tons, in the announcement post they point out that valued at API prices, the median researcher at OpenAI uses $600 in tokens per day, while the 90th percentile of researchers use $7,000 of tokens per day.

他们几乎轻描淡写地写道：“随着编程代理承担更长、更复杂的任务，持续使用的成本就变得更加重要。”

Almost understating it, they write, "As coding agents take on longer and more demanding tasks, the cost of sustained use matters more."

结论是：“GPT-6 Sol 和 Luna，”他们写道，“结合了强大的编程性能和更低的 API 价格，给开发者更多迭代空间，也让团队更有信心去让 Codex 承担更宏大的任务。”

The upshot, "GPT-6 Sol and Luna," they write, "combine strong coding performance with lower API prices, giving developers more room to iterate and teams the confidence to be more ambitious about what they ask Codex to take on."

那外部基准测试呢？

But what about external benchmarks?

即使是在乎基准测试的人，也都知道内部数据不能全信。

Even for those folks who care about benchmarks, everyone assumes you have to take any internal measures with a grain of salt.

新模型一出来，大家首先会去看像Artificial Analysis这样的独立来源。

And one of the first places that people look when a new model comes out is independent sources like Artificial Analysis.

在他们新的V3指数上，GPT-5.6 Luna和GPT-6 Luna的总分基本一样，都是37，而GPT-6 Sol稍微超过了GPT-5.6 Sol。

On their new V3 index, GPT-5.6 Luna and GPT-6 Luna are basically scored the same at 37 overall, while GPT-6 Sol slightly beats GPT-5.6 Sol.

不过，这两个模型达到这个分数，单位智能的成本更低。

That said, both of those models get those scores at a lower cost per intelligence.

实际上，Artificial Analysis的第一句话就是：“GPT-6 Sol和Luna把成本效率推到了新高度，成本比GPT-5.6 Sol和Luna降低了一半。”

Indeed, Artificial Analysis' first sentence is, "GPT-6 Sol and Luna push the cost efficiency frontier by halving cost relative to GPT-5.6 Sol and Luna."

另一个亮点是，Artificial Analysis发现这两个模型的幻觉问题都明显减少了。

In another positive highlight, Artificial Analysis found that both models saw a significant reduction in hallucination.

Zapier也在一系列真实工作流程中测试了GPT-6 Sol。

Zapier also tested GPT-6 Sol on a slate of real workflows.

他们发现6 Sol得分33.2%，而GPT-5.6 Sol是28.77%，而且价格大约是5.6的一半。

They found that 6 Sol scored 33.2% compared to GPT-5.6 Sol's 28.77%, and did so at roughly half the price of 5.6.

他们的总结很简单。

Their summation was pretty simple.

“把任何使用5.6的工作流都升级了吧。

"Upgrade any of your workflows that are using 5.6.

它更便宜，也更好。

It's cheaper and better.

这种组合很少见。”

A rare combo."

不过，Opus 5.5就完全是另一回事了。

Opus 5.5, however, was a different kettle of fish entirely.

Opus 5.5之所以引人注目，是因为它在智能指数上以整整5分的优势碾压了其他所有模型，拿下了新的最高分。

Opus 5.5 was notable because it absolutely thwomped everything else to get the new top score on the intelligence index by five whole points.

之前Claude Fable 5.1和GPT-6 Astra以53分并列榜首，但Claude Opus 5.5直接跳到了58分。

Claude Fable 5.1 and GPT-6 Astra were previously holding it down at the top with scores of 53, but Claude Opus 5.5 jumped all the way to 58.

虽然它的成本和效率提升可能不如GPT-6 Sol那么明显，但Artificial Analysis确实指出，这次性能提升还伴随着20%的降价。

And while its cost and efficiency gains might not have been as much as GPT-6 Sol, Artificial Analysis did note that this increase in performance also came with a 20% price cut.

说实话，虽然我一直尽量让这次分析保持对齐和一致，但看看大家的初步反应，毫无疑问这些模型里哪个主导了讨论。

And honestly, while I've been trying to keep this analysis largely aligned and equivalent so far, watching people's first responses, there is absolutely no doubt which model among these dominated the discourse.

Matthew Berman写道：“Opus 5.5是地球上绝对最好的模型，而且比Astra和Fable还便宜，这我可真没想到。”

Matthew Berman wrote, "Opus 5.5 being the absolute best model on the planet and being cheaper than Astra and Fable was not on my bingo card."

Chubby Kiminismus在X上写道：“我越用越喜欢Opus 5.5。”

Chubby Kiminismus on X writes, "The more I use, the more I fall in love with Opus 5.5.

它太好了，快得多，而且简洁得多。

It's so good and so much faster and so much less verbose.

它满足了我所有的期望。

It's everything I could have asked for.

感觉就像我心爱的Opus 4.6回来了，但更好，还改名为5.5。”

It's as if my beloved Opus 4.6 came back but better and rebranded as 5.5."

现在我们再深入聊聊这种喜爱，但人们具体在哪些方面测试Opus 5.5呢？

Now we'll get more into that love, but what were some of the specific areas that people were testing Opus 5.5 on?

Aaron Levy 和 Brock's 的团队把它引入到各种复杂的、处理非结构化数据的企业知识工作场景中，不仅看到了性能的提升，还获得了显著的效率增益。

Aaron Levy and the team at Brock's brought it into, quote, "a variety of complex enterprise knowledge work tasks dealing with unstructured data," and saw not just improvements in performance, but major efficiency gains as well.

跟 Opus 5 相比，他们发现 token 使用量减少了 63%，冗余度降低了 42%，速度提升了 30%。

Versus Opus 5, they found 63% fewer tokens used, 42% less verbosity, and 30% faster.

在金融服务任务上，他们发现任务准确率提高了 39%。

On financial services tasks, they found an increase of 39% task accuracy.

在云成本分析和技术应用场景下，他们发现任务准确率提高了 65%。

On cloud cost analysis and technology use cases, they found an increase of 65% task accuracy.

在涉及消费产品和临床诊断的应用场景中，他们发现任务准确率分别提升了17%和15%。

And in use cases related to consumer products and clinical diagnostics, they found an increase of 17% and 15% task accuracy respectively.

如果你在企业里负责AI相关事务，你应该已经意识到，现在的差距不在于能力，而在于落地执行。

If you're leading AI inside an enterprise, you already know that the gap right now isn't capability, but execution.

正因如此，KPMG的《你可以用AI》节目带着新一季回归了，这一季邀请了Emma的Sarojit Chatterjee、Writer的May Habib、McKesson的首席信息官Ellery Fisher等领袖人物，围绕实际执行展开对话。

That's why KPMG's You Can with AI is back with a new season featuring conversations with leaders like Sarojit Chatterjee of Emma, May Habib of Writer, McKesson CIO Ellery Fisher, and others focused on practical execution.

看看哪些做法有效，哪些行不通，以及要在战略、数据准备、治理、团队建设和价值创造这几个方面从试点走向规模化落地，到底需要付出什么。

What's working, what's not, and what it actually takes to move from pilots to real scaled impact across strategy, data readiness, governance, workforce, and value.

当然，节目由我和纳撒尼尔·惠特莫尔共同主持。

And of course, it's co-hosted by me, Nathaniel Whittemore.

快去 www.kpmg.us/aipodcasts 收听并订阅吧。

Go listen and subscribe at www.kpmg.us/aipodcasts.

网址是 www.kpmg.us/aipodcasts。

That's www.kpmg.us/aipodcasts.

下面我们来聊聊，为什么大多数传统系统的现代化改造都会失败。

Here's why most legacy modernization projects fail.

靠 AI 干活，根本没法大规模理解代码库。

The AI doing the work can't understand code bases at scale.

它只能看到一小部分上下文，看看语法，却漏掉了全球应用生态里多年来积累的那些决策。

It sees a small slice of context, examines syntax, and misses years of decisions distributed across the global application ecosystem.

Blitzy 解决这个问题的方式，跟它解决所有问题一样。

Blitzy solves this the way it solves everything.

在迁移开始之前，Blitzy 的代理就会先扎根到你的代码里，把整个遗留系统逆向工程成一个持久的知识图谱。

Grounded in your code before any migration begins, Blitzy's agents reverse engineer the entire legacy system into a persistent knowledge graph.

每一个依赖、每一个约束、每一段曾经只存在于某个工程师脑子里的隐性知识。

Every dependency, every constraint, every piece of tribal knowledge that used to live in one engineer's head.

基于这种理解，Blitzy 自主执行语言迁移、框架升级以及从单体架构到微服务的转型，并且全程进行端到端验证。

From that understanding, Blitzy autonomously executes language migrations, framework upgrades, and monolith to microservices transformations, all validated end to end.

Blitzy 的一位客户，用编码智能体在16周内完成了价值1000万美元的保险单体系统现代化改造，而传统基线需要137周。

One Blitzy customer modernized a $10 million monolithic insurance stack in 16 weeks against a 137-week baseline with coding agents.

这相当于压缩了9倍的时间。

That's 9x compression.

在加速你的路线图的同时，清除技术债务。

Retire technical debt while accelerating your roadmap.

访问 blitzy.com 了解更多。

See how at blitzy.com.

网址是 B-L-I-T-Z-Y 点 com。

That's B-L-I-T-Z-Y dot com.

接下来要说一个残酷的事实。

Here's a harsh truth.

你们公司可能花了几千甚至几百万美元买AI工具，但实际利用率低得可怜。

Your company is probably spending thousands or millions of dollars on AI tools that are being massively underutilized.

一半的公司都买了AI工具，但真正用来创造业务价值的只有12%。

Half of companies have AI tools, but only 12% use them for business value.

大多数员工还在拿AI来整理会议纪要。

Most employees are still using AI to summarize meeting notes.

如果你负责公司的AI落地工作，那你一定要试试Section。

If you're the one responsible for AI adoption at your company, you need Section.

Section是一个平台，能帮你在全公司范围内管理AI转型。

Section is a platform that helps you manage AI transformation across your entire organization.

它会指导员工处理实际应用场景，追踪谁在用AI创造业务价值，并让你清楚看到AI在哪些地方产生了价值，哪些地方没有。

It coaches employees on real use cases, tracks who's using AI for business impact, and shows you exactly where AI is and isn't creating value.

这样一来，你就能从单纯部署工具，转变为推动可衡量的AI价值。

The result, you go from rolling out tools to driving measurable AI value.

你的员工不再只是做会议摘要，而是去解决真正的业务问题，而且你能证明投资回报率。

Your employees move from meeting summaries to solving actual business problems, and you can prove the ROI.

别再猜你的AI投资有没有效果了。

Stop guessing if your AI investment is working.

去sectionai.com看看Section吧。

Check out Section at sectionai.com.

就是S-E-C-T-I-O-N-A-I点com。

That's S-E-C-T-I-O-N-A-I dot com.

本期《AI每日简报》由Hyperagent赞助播出，你可以通过它管理整个代理团队。

This episode of the AI Daily Brief is brought to you by Hyperagent, where you run fleets of agents your team can manage together.

别再依赖本地代理和聊天工作流，等着你的笔记本电脑来提示了。

Forget local agents and chat workflows waiting on your laptop to be prompted.

Hyperagent在云端部署全天候运行的代理，在你团队已经使用的工具中完成实际工作。

Hyperagent deploys always-on agents in the cloud, doing real work across the tools your team already uses.

营销代理能把竞争对手的动态变成落地页。

Marketing agents turn competitor moves into landing pages.

销售代理能丰富潜在客户信息、起草邮件，并更新CRM系统。

Sales agents enrich leads, draft emails, and updates the CRM.

运营代理负责处理文书工作和跟踪预算。

Ops agent chases the paperwork and tracks the budget.

每个代理都能访问共享上下文，并遵守你关于范围和审批的规定。

Every agent has access to shared context and follows your rules about scope and approvals.

是时候让你的代理像队友一样了。

It's time you had agents that feel like teammates.

去Hyperagent雇佣你的代理吧。

Hire yours at Hyperagent.

在hyperagent.com/aidailybrief获取100美元积分。

Get $100 in credits at hyperagent.com/aidailybrief.

现在，说到社交媒体演示，和最近几款模型一样，大量帖子都是高度视觉化的内容，比如图形、3D渲染和游戏设计，这些在网上看起来很有趣，但未必能真实反映大多数人实际会使用这款模型的场景。

Now, when it came to the social media demos, as has been the case with the last several models, a huge amount of the posts were highly visual things like graphics, 3D renderings, and game design, which are interesting to look at on the internet, if not necessarily super reflective of the type of use cases that most people are going to actually have for this model.

不过，考虑到3D图形、视觉效果和Blender在GPT-6 Astra发布中占了很大比重，Opus 5.5的很多演示基本上就是在这些使用场景上碾压了那款模型。

Still, given how much 3D graphics and visuals and Blender were a big part of the story of the GPT-6 Astra release, a lot of the demos of Opus 5.5 basically have it mogging that model on exactly that type of use.

而且，鉴于我就在一两周前，或者不管什么时候我们拿到GPT-6 Astra时做的分析，是关于这些能力如何促使我们以更广阔的视角思考AI带来的机遇，值得注意的是，Opus 5.5至少初步看来，在这些能力集上似乎又是一个飞跃。

And given that my analysis just a week or two ago, or whenever it was that we got GPT-6 Astra, was about how much those sort of capabilities asked us to think in more expansive terms about the opportunities that AI opened up, it's worth noting that Opus 5.5, at least initially, seems to be another jump in those capability sets as well.

Anthropic团队的Alex Albert展示了如何用Opus 5.5配合Blender，通过一个提示词就能制作黏土动画。

Alex Albert from the Anthropic team showed Opus 5.5 using Blender to make claymations with a single prompt.

Chase Leen制作了一款互动式的珊瑚礁壁纸。

Chase Leen made an interactive coral reef wallpaper.

还有人演示了模型有多强大，能够直接生成代码来构建真实的世界视觉画面，而不是依赖图像生成模型。

And some people showed how the models were powerful enough to use generated code rather than image models to create real visual representations of the world.

Peter Yang分享了一段看起来像金门大桥的视频，但他指出这其实完全是由Opus 5.5通过代码生成的。

Peter Yang shares what looks like a video of the Golden Gate Bridge, but notes that it's actually entirely generated using code by Opus 5.5.

他补充说，"根据我的测试，我觉得 Opus 在搭建 3D 场景方面和 Astra 一样厉害。"

He adds, "From my testing, I think Opus is just as good at building 3D scenes as Astra."

Alex Albert 也做了类似的操作，只用 Opus 5.5 的代码和 Blender，没有生成任何图片，就还原了 1906 年地震前旧金山市场街的历史真实风貌。

Alex Albert again did something similar, visualizing a historically accurate San Francisco Market Street in 1906 pre-earthquake using only Opus 5.5 code and Blender, no image generation.

Dan Wood 表示，Opus 5.5 引号"在这些 3D 生成世界的表现上，绝对把其他模型都甩在身后引号，并补充说这是引号"空间感知能力的一次惊人飞跃引号。"

Dan Wood says that Opus 5.5, quote, "Absolutely frame mogs every other model on these sort of 3D generated worlds," adding that it is a, quote, "Astounding leap in spatial awareness."

Anthropic 的 Jake Eaton 发了一条超级火的帖子，分享了好几张看起来像是画作的作品，他说："过去几个月，我一直在让咱们的模型画画。"

In a wildly viral post, Jake Eaton from Anthropic shared a number of different what appear to be paintings and said, "For the past few months, I've been asking our models to paint.

Opus 5.5 在模仿不同风格方面非常擅长。

Opus 5.5 is very skilled at emulating different styles.

这里的每一张图都是一个 Python 程序逐像素生成的。

Every image here is a Python program generated pixel by pixel.

没有用到图像模型，也没有现成的艺术软件。

There is no image model and no off-the-shelf art software.

相反，它用了大约 7500 行代码，通过标准库来模仿不同的画笔风格。

Instead, it's about 7,500 lines of code using standard libraries to emulate different brush styles.

这些代理没有参考任何图片，只根据它们对每位画家的了解来创作。"

The agents don't use any pictures as reference, instead working only from what they know about each painter."

当然，正是在这里，你可以看到这些能力——这些能力在社交媒体上看起来可能很酷，但对大多数人来说不一定那么实用——也许开始变得更有实际意义了。

And certainly, this is where you can see these capabilities, which might otherwise be cool to look at on social media, but not necessarily all that relevant for most of us, maybe start to become a little bit more relevant.

如果 Opus 5.5 的高级编码能力足以像 Pack 的这条帖子那样，一次性生成一段 30 秒的营销动画，那么毫无疑问，突然之间就会涌现出许多你以前可能没想到的用例。

If the advanced coding capabilities of Opus 5.5 are enough that it can, as in the case of this post by Pack, generate a one-shot, 30-second long marketing animation, well, certainly all of a sudden a bunch of use cases open up that you might not have thought of before.

Claude Code 团队的 Tariq 让 Opus 对他的个人网站做了一系列重新设计，然后用所有的迭代版本制作了一个预告片，乍一看，它似乎做得非常出色。

Tariq from the Claude Code team asked Opus to do a bunch of redesigns on his personal website and then make a trailer with all of its iterations, and at first glance, it seems to have done a great job.

顺便说一句，这个用例我也在 Opus 5.5 上试过，我对它对我 AI Daily Brief 网站的分析和评价非常满意。

For what it's worth, this is one use case that I tried on Opus 5.5 as well, and I was definitely pleased with its analysis and review of my AI Daily Brief website.

它思路清晰，表达简洁，而且非常实用。

It was clear, had good concision of thought, and was very practical.

不过我要说，对于那些只想用好坏来评判的人来说，它并不一定比 Fable 5.1 的分析更好，只是风格不同。

Although I will say for those who are just trying to live in the world of strictly betters or strictly worses, it wasn't that it was necessarily strictly better than Fable 5.1's analysis, it was just different.

然而，有一个领域几乎所有人都认为 Opus 5.5 确实更胜一筹，那就是写作。

And yet, one area where almost everyone seems to be in agreement that Opus 5.5 is indeed strictly better is writing.

Anthropic 的 Sholto Douglas 转发了他们的公告，并写道：“还有个重要消息，我们修复了写作问题。”

Anthropic's Sholto Douglas reposted their announcement and wrote, "Also important news, we fixed the writing."

那这意味着什么呢？

So what does that mean?

嗯，部分原因是，正如 Theo 用大写字母说的，他们去掉了破折号。

Well, part of it is, as Theo put it in all caps, they removed the em dashes.

但 Anthropic 官方账号对此的实际说法是：“Opus 5.5 的沟通更自然，解决了我们在 Opus 5 上收到的一些最常见的反馈。”

But what the Anthropic account actually said about this is, "Opus 5.5 communicates more naturally, addressing some of the most common feedback we heard on Opus 5.

它把最重要的信息放在前面，并且遵循你给它的写作规则，这让长时间的对话更容易跟上。

It puts the most important information up front and follows the writing rules you give it, which makes long sessions easier to follow."

Every 团队拥有测试 AI 写作的最佳基准和流程之一，他们写道：“Opus 5.5 是我们见过的 Anthropic 或 OpenAI 模型中最易读的文本。”

The team at Every, who have one of the best benchmarks and processes for testing AI writing, wrote, "Opus 5.5 produces the most readable prose we've seen from an Anthropic or OpenAI model."

实际上，他们预见到了对话中稍后会更多提到的一点，写道，虽然 Opus 5 让他们的作者把 Claude 模型弃之不用，但 5.5，“让我们想让 Claude 回到房间里。”

In fact, presaging a point that would come up a lot more later in the conversation, they wrote that while Opus 5 had made their writers kick Claude models to the curb, 5.5, quote, "Makes us want Claude back in the room."

他们继续说：“它对反馈反应良好，能基于你提供的材料进行构建，并且在 Opus 5 不会解释的地方解释其选择。”

They continue, "It responds well to feedback, builds on the material you give it, and explains its choices where Opus 5 wouldn't."

重要的是，事实证明这不仅是对写作输出的改进，而且是对实际过程的改进。

And importantly, it turns out that this isn't just an improvement on the output of writing, but an improvement on the actual process itself.

他们写道，Opus 5.5 写起来很愉快。

They write that Opus 5.5 is a pleasure to write with.

它接受反馈而不争吵，并且能基于材料进行构建，而不是只是把材料整理得更整洁地还给你。

It takes feedback without a fight and builds on material instead of handing it back tidier.

实际上，团队写道：“最大的升级甚至不完全在于写作本身。”

In fact, the team writes, "The biggest upgrade is not even exactly about the writing.

这关乎模型给人的感觉和氛围。

It's about the feel and vibe of the model.

正如Every团队所说，Anthropic修复了Opus的个性问题。

As the Every team puts it, Anthropic fixed Opus's personality.

它不再是个固执的小混蛋了。

It's not an obstinate little turd anymore."

而且很多人都有这种体验。

And this is an experience that many have had.

AI开发者McKay Wrigley写道：“Opus 5.5具备了我们都迫切希望回归的Opus 4.6的个性，以及Fable 5.1的智慧和品味。”

AI builder McKay Wrigley writes, "Opus 5.5 equals the personality of Opus 4.6 that we all desperately wanted back, and the intelligence and taste of Fable 5.1."

当然，Anthropic自己也清楚这一点。

And certainly, Anthropic knew it.

Anthropic的Nat McCallis发帖说：“Opus 5.5比Opus 5好太多太多了。

Nat McCallis from Anthropic posted, "Opus 5.5 is way, way, way better than Opus 5.

那个模型真是抱歉了。”

Sorry about that model.

大家不妨试试这个。

Please try this one."

那么，目前有人在使用 Opus 5.5 时遇到问题吗？

So have people had any issues yet with Opus 5.5?

确实有一个问题比较突出，在某些特定场景下，这基本上就是不可行的。

There is one that stands out, and that for certain use cases will be effectively a non-starter.

X 平台上的 Chief of Staff 写道："我把重置的使用额度全烧光了，用来把 Opus 5.5 和 Fable 5.1 做基准测试，主要是针对复杂的开放式法律工作，比如起草面向客户和法庭的文件、合同审阅与修订，以及法律和案件相关的研究。"

Chief of Staff on X wrote, "I burned all my usage reset benchmarking Opus 5.5 against Fable 5.1 for complex open-ended legal work, i.e. drafting client and court-ready documents, draft and contract redlining, legal and matter-related research.

据我观察，Opus 5.5 表现差了不少，主要是安全拦截变多了，而且我觉得它在算力分配上也不太合理。

As far as I can tell, Opus 5.5 is much worse due to increased safety rejections and what I'd guess you would call poor effort budgeting."

Simon Smith 写道："Opus 5.5 看着不错，但作为一家生命科学商业化公司，我有点担心。"

Simon Smith writes, "Opus 5.5 looks good, but as a life science commercialization company, I'm a bit concerned by this."

"因为 Opus 5.5 在生物和网络安全方面跟 Claude Mythos 5.1 差不多，所以我们打算给它加上类似 Claude Fable 5.1 的安全限制。"

Quote, "Because Opus 5.5 is comparable to Claude Mythos 5.1 in biology and cybersecurity, we're deploying it with safeguards similar to those on Claude Fable 5.1."

Simon 接着说："我们在用 Fable 处理某些任务时遇到了问题，就是因为这些限制。"

Simon continues, "We faced issues using Fable for some tasks because of these guardrails.

我们已经申请了Anthropic的生命科学验证项目，但还没获批。

We've applied for Anthropic's life science verification program but haven't yet been approved.

之前Fable因为生物学防护措施拒绝请求时，我们用的是Opus，但现在看来Opus这边也不行了。

We had used Opus when Fable refused a request because of Fable's biology guardrails, but now it seems like this won't be possible anymore with Opus either."

虽然不是所有地方都这样，但我确实看到不少人在抱怨这些过度防护的措施。不幸的是，这种事一旦撞上你的使用场景，基本上这个模型就废了。

It's not everywhere, but I certainly have seen a number of people complaining about these overzealous safeguards, which is unfortunately the type of thing where if that hits your use case, it basically makes the model null and void.

那这一大波发布之后，大家有什么重要的收获吗？

So what are some of the big takeaways for folks after this huge slate of releases?

Aaron Levy写道：“今天真是AI界疯狂的一天。

Aaron Levy writes, "What an insane day in AI.

前沿模型的价格大幅下降，Opus 5.5降价了，现在GPT-6 Sol和Luna又把token价格砍了50%。

The frontier models just became substantially cheaper with the Opus 5.5 price cuts, and now with GPT-6 Sol and Luna dropping token prices by 50%.

AI中每项任务成本下降的速度，在同类基础上比较，是历史上任何其他技术都无法比拟的。每次AI成本下降，你能部署agent的使用场景就会大幅增加。

The rate at which the cost per task on a like for like basis drops in AI is unlike any other type of technology in history, and every time the cost of AI drops, the use cases you can deploy agents against dramatically increase.

这就是杰文斯悖论在agent上的体现。”

This is Jayvon's paradox applied to agents."

Aaron 继续预测说：“这些改进将直接推动 AI 在经济中更广泛地普及，因为我们可以用代理来处理所有数据、扫描代码中的安全问题、阅读所有日志数据来做决策，还有代理集群和工作流，等等。”

Making a bit of a prediction, Aaron continues, "These improvements will directly lead to broader diffusion of AI in the economy, as we can use agents to process all of our data, scan our code for security issues, read through all log data to make decisions, have agent swarms and workflows, and much more.

“Token 的成本直接决定了这些用例能否大规模落地。”

The cost of tokens is directly correlated to these use cases being opened up at scale."

为了给这个观点提供研究支持，Epic AI Research 昨天也发布数据称：“AI 的成本下降速度比历史上任何其他变革性技术都要快。”

And putting some research heft behind that, Epic AI Research yesterday also dropped data arguing that, quote, "AI is getting cheaper more quickly than any other transformative tech in history.

“在同等性能水平下，自 2023 年以来，成本每季度下降约 47%。”

At a given level of performance, cost has fallen around 47% per quarter since 2023."

他们指出，这比 DNA 测序快 4 倍，比计算快 6 倍，比锂电池快 18 倍，而且至少到 1973 年，比电力快 54 倍。

They point out that that's 4x faster than DNA sequencing, 6x faster than compute, 18x faster than lithium batteries, and at least up to 1973, 54 times faster than electricity.

现在，说到正面比较，你会发现对于这些模型中哪些总体上最有价值，各方观点不一。

Now, when it comes to head-to-head analysis, you can find people arguing all sides when it comes to which of these models are the most valuable in aggregate.

ToYourTaxes 写道：“更便宜、更好的 Sol 和 Luna。”

ToYourTaxes writes, "Cheaper, better Sol and Luna.

“我猜它们天生就是为当 Astra 的跟班而训练的。”

I assume they're natively trained for being Astra's henchmen.

这让OpenAI的价值主张变得更有竞争力了。

This makes OpenAI's value proposition considerably stronger.

我觉得Opus 5.5在成本和多智能体方面的优势，基本上被Astra加上Sol和Luna抵消了。

I suspect that Opus 5.5 wins on cost and multi-agent are effectively negated with Astra plus Sol plus Luna."

也有很多人就是喜欢享受这些不同模型各自带来的独特价值。

And there are also many folks just happy to enjoy the unique values that each of these different models provided on their own terms.

在回答谁赢了这个问题上，Chubby写道：“OpenAI和Anthropic都给了我们很多值得兴奋的东西。”

In answering the question of whether there was a winner, Chubby writes, "Both OpenAI and Anthropic gave us plenty to be excited about."

如果你对OpenAI的战略有任何疑问，Sam Altman已经说得再清楚不过了。

And if you have any question about the strategy that OpenAI is pursuing, Sam Altman could not have made it any more clear.

他写道：“我们希望OpenAI API在每个价格点上都有最好的模型，并且在每个模态上都是最好的，包括文本、代码、图像、视频等等。”

He wrote, "We want the OpenAI API to feature the best model at every price point and to be the best at every modality, text, code, image, video, et cetera.

然后我们都希望你能想出好主意，把它们做出来，成为快乐的用户。”

And then we all want you to come up with great ideas and build them and get to be happy users."

再说一次，记得他在另一篇文章里写过：“我们希望人们能大量使用AI。”

Again, remember that he had written in a separate post, "We want people to be able to use tons of AI.

这对我们探索眼前这场新文艺复兴来说很重要。

It is important to being able to explore this new renaissance in front of us."

我觉得从很多方面来看，你可以把OpenAI这一系列发布看作是他们一直在追求的战略的延续，那就是在成本和效率上非常激进，并且为这样一个时代构建一套模型组合——在这个时代，人们正从单一模型转向复杂的模型架构，把任务和合适的智能类型匹配起来。

I think in many ways, you can view this set of releases from OpenAI as a continuation of the strategy that they've clearly been pursuing for a while now, which is be really aggressive about focusing on cost and efficiencies and building a model slate for an era in which people are moving away from a single model towards complex model architectures that actually match tasks with the right type of intelligence.

另一方面，Claude感觉像是为了重新夺回一些势头而发布的，如果真是这样，那他们做得相当成功。

Claude, on the other hand, feels like it was out to reclaim some momentum, and if that is the case, they've done so pretty successfully.

Every是这么说的。

Here's how Every put it.

“Sol感觉像是S级的iPhone发布。

"Sol feels like an S-class iPhone release.

它会以大约五分之一的价格给你带来Astra的大部分能力。

It will give you much of Astra's power at about a fifth of the price.

Opus 5.5才是更大的惊喜。

Opus 5.5 is the bigger surprise.

5.5正在诱惑我们团队里的几个人从Codex切换回来。”

5.5 is tempting a few people on our team to switch back from Codex.

如果你已经在用Claude生态，你会喜欢这个模型的；如果你是Codex用户，也值得一看，尤其是处理高端编程任务的时候。

You'll love this model if you're already in the Claude ecosystem, and if you're a Codex user, it's worth a look, especially for your top-end coding tasks."

现在，从某种程度上说，这只是我之前提到的那个分歧的延续，我们已经看了好几代模型了。

Now, in some ways, this is just a continuation of the divide which I mentioned before and which we've seen for several model iterations now.

Every的Dan在文章结尾写道：“如果你想要一个快速、实惠的日常工具，用来阅读、写作和处理事务，你会喜欢Sol 6的。

Dan from Every concludes that article, "You'll like Sol 6 if you want a fast, affordable daily driver for reading, writing, and getting things done.

它是我一直想用的模型。

It's the model I keep reaching for.

如果你愿意花更多钱，在雄心勃勃的编程和视觉项目上获得更强的性能，你会喜欢Opus 5.5的。

You'll like Opus 5.5 if you're willing to pay more for stronger performance on ambitious coding and visual projects.

在我们的测试中，它的最佳表现超过了Sol，足以让团队里一些人重新回到Claude的怀抱。”

Its best work surpassed Sol in our tests, enough to pull some of our team back towards Claude."

而且，把人们拉回Claude这个说法，确实是很多人的真实感受。

And the pulled people back to Claude is a real take that I'm seeing from a number of different people.

Yu Chen Jin写道：Claude回来了。

Yu Chen Jin writes Claude is back.

是时候让我再次打开Claude Code了，我已经一个月没碰它了。

Time for me to open Claude Code again after ignoring it for a month.

而且，时间越久，我看到人们对Opus 5.5的评价就越正面。

And if anything, the more time that goes on, the more positive I'm seeing people get on Opus 5.5.

今天早上，前投资人转行做AI的Jeffrey Emanuel又发话了：“天哪，Opus 5.5简直令人叹为观止。

This morning, former investor turned AI builder, Jeffrey Emanuel, chimed back in, "My God, Opus 5.5 is breathtaking.

它处理那些极其棘手的问题就像小菜一碟。

It's just grinding through incredibly tricky stuff like it's nothing.

它找到了一些Fable和Astra几周都没发现的bug和问题，展现出了我从未见过的自主性和决心。”

Finding bugs and problems that eluded Fable and Astra for weeks in some cases, and showing a level of agency and resolve I haven't seen before."

就我个人而言，我打算再多花点时间用用这些模型，然后会在本周晚些时候做一期更偏实操的节目，教大家如何充分利用这两个模型。

Now, for me personally, I'm going to save a bit of my personal analysis for a little bit more time with the models, which will come together in more of a how-to episode for getting the most out of both of these models later in the week.

不过，我确实觉得有几个有趣的观察值得在最后提一下。

That said, I do think that there are a few interesting observations that are worth closing on.

顺便说一句，我讲解时用的这个视觉图是GPT-6 Sol生成的，它加了不少我原本没有的文字，但我留着它，也许可以当作一个展示特定模型怪癖的例子。

By the way, this visual that I'm talking over was created by GPT-6 Sol, which added quite a bit of text that I didn't have in there, but I left it perhaps as an example of particular model quirks.

第一个观察是，即使是那些更果断切换、自称只在乎性能的早期用户，也肯定有性格偏好。

Observation number one is that even early adopters who are more ruthless about switching and who profess to really only care about performance definitely have personality preferences.

Opus 5.5 带来的兴奋点，很大程度上不是‘哇，它做到了AI以前做不到的事’，而是‘天哪，用起来这么舒服，好久没这种感觉了’。

So much of the excitement around Opus 5.5 is not about, wow, it does this thing which AI has never been able to do before, but about, oh my gosh, it is such a pleasure to use this in a way that it hasn't been for some time.

我觉得这一点其实特别重要。

I actually think that this is extremely important.

当 ChatGPT-4o 被下架引发那么大争议时，有人倾向于把它当成只是那些对奉承模型着迷的普通用户的现象。

When there was such a big dust-up around ChatGPT-4o being deprecated, there was a temptation to treat it like a phenomenon exclusively for crazy normies who had gotten obsessed with the sycophantic model.

但随着时间的推移，尽管AI对我们有些人来说可能是个工具，它显然是一种非常不同的工具。

But as time has gone on, for as much as AI might be a tool to some of us, it's clearly a very different type of tool.

这个工具至少近似于有性格，而不同版本的工具性格，以及使用起来是否更愉快，实际上会影响我们能从中获得多少。

It's a tool that at least approximates having a personality, and different versions of that tool's personalities, and more or less enjoyable to use, actually has implications for how much we can get out of them.

简而言之，对于大语言模型来说，性格就是用户体验，我们应该确保这样对待它。

In short, when it comes to LLMs, personality is UX, and we should make sure we're treating it as such.

第二个观察是，虽然如果你只看热度和兴奋度，肯定会把 Opus 5.5 评为赢家，但大语言模型的竞争其实已经不再只关乎一件事了。

Observation number two, while it is definitely the case that if you were just looking to declare one winner from the standpoint of buzz and excitement, it would be Opus 5.5, the battle of LLMs is not really about one thing anymore.

这涉及到两个方面：一是要弄清楚需要什么样的工具组合，二是要为这个组合中的每个位置去竞争。

It's about, one, figuring out what the stack of options needs to be, and then two, battling for each slot in that stack.

换句话说，即使是最喜欢 Opus 5.5 的人，也不会说每个用例都必须用它。

In other words, even the people who loved Opus 5.5 the most aren't saying you have to use this for absolutely every single use case.

而且在很多方面，我甚至觉得，虽然乍一看 Opus 5.5 和 GPT-6 Sol 似乎是直接竞争对手，但它们的开发者对它们的思考方式其实有点不同。

And in many ways, I'd even argue that although at first glance, Opus 5.5 and GPT-6 Sol seem to be direct competitors, the way that their respective builders are thinking about them is a little bit different.

在这个阶段，重要的是个人和公司能够理解自己的全部需求，以及不同的模型和工具如何组合起来，最好地满足这些需求，包括性能、能力、成本效益和速度。

What matters at this stage is individuals and companies being able to understand their complete set of needs and how different models and harnesses combine to best fill those needs with respect to performance, capability, cost efficiency, and speed.

第三个观察是，我们现在确实处于一个无法再把模型和工具完全分开的时代，至少不能完全分开。

A third observation is that we're definitely in the era where you can't really separate models from harnesses anymore, at least not fully.

比如，很多现在已经完全投入 Codex 生态系统的人，即使 Opus 5.5 好得多，对他们来说也无所谓。

For a lot of folks who are fully invested, for example, in the Codex ecosystem now, it doesn't matter if Opus 5.5 is much better.

只要他们的 OpenAI 选项足够接近，他们就不会完全离开那个工具。

As long as their OpenAI options are close enough, they're not going to fully shift out of that harness.

Prime Intellect 的 Will Brown 写道：“我每周都在换桌面代理，这已经失控了。

Will Brown from Prime Intellect wrote, "I keep switching desktop agents every week, and it's getting out of control.

Codex 没有 Fable 5.1，Claude 也没有 Astra。

Codex didn't have Fable 5.1, Claude didn't have Astra.

AMP 之前两个都有，但现在没了 Opus 5.5，所以我又回到 Claude 了。

AMP had them both, but now doesn't have Opus 5.5, so I'm back to Claude.

自己搭建维护起来太烦人了。"

Self build is annoying to maintain."

这期节目里我们没怎么聊到框架（harness）的事，但大概过一周左右，真正的考验就会见分晓：看看大家是会因为这些模型的变化而真正改变使用习惯，还是说大家都试了一圈，知道自己喜欢什么后，又回到了原来就用的那个生态体系里——这在很大程度上是由他们存放上下文、工具和规则等内容的框架决定的。

We didn't talk much about harnesses in this episode, but give it about a week, and the real proof in the pudding will be whether anyone has actually switched their overall behavior because of any of these model changes, or whether everyone tested things, found out what they liked, and just went back to the ecosystem that they were already using before, dictated in large part by the harness where they store their context, their tools, their rules, et cetera.

第四个观察点其实跟Anthropic或OpenAI关系不大，而是关于一个事实：在这些模型发布的同时，大家热议的另一大焦点，我认为是一个真正获得巨大成功的AI产品——使用者根本不在乎也不了解它底层用的是哪个模型。

A fourth observation is actually not so much about either Anthropic or OpenAI, but about the fact that as these models were being released, the other big buzzy thing that people were talking about was, I would argue, the first AI product that is seeing a lot of traction, where the people using it genuinely don't know or don't care about the model underneath.

我说的当然是Muse。从Meta首席AI官Alexander Wang的发文就能看出来，Meta团队对这款产品受到的欢迎和反响感到非常兴奋。

I'm talking, of course, about Muse, which you can tell just from the posting from Meta's chief AI officer, Alexander Wang, the team over there at Meta is very excited about the uptake and reception for.

如果模型真的被逐渐抽象化、隐藏起来，这会怎么改变我们目前关于新模型的讨论呢？

How will it change this sort of new model conversation if the models actually start to get abstracted away?

这个问题我们聊过很久了，但到目前为止，在实际应用中还没真正显现出来。

That's something we've talked about for a long time, but which so far hasn't really played out in practice.

但谁知道呢？

But who knows?

也许我们现在正处于AI的产品时代，而不是模型时代，模型真的会消失在系统里。

Maybe we're at the beginning of the product rather than the model era of AI, in which the models really will disappear into the system.

今天第五个也是最后一个观察，就是这一切对前沿节奏意味着什么。

The fifth and final observation for today is what this all says about pacing the frontier.

在X上，一个常见的段子能让你获得大量转发和点赞，就是人们睁着疲惫的眼睛，看着两天内四个主要模型发布，然后问某种版本的问题：“这就是前沿节奏？”

One of the standard jokes that could get you a bunch of reposts and likes on X was people staring with bleary eyes at four major model releases in two days and asking some version of, "This is pacing the frontier?"

但我认为Theo说得对，他写道：“大胆说一句，这就是节奏的样子。”

But I think Theo had the right of this when he wrote, "Hot take, this is what pacing looks like.

今天的发布没有一个达到Astra或Fable的水平。

None of today's releases were Astra or Fable tier.

这是故意的。

This is intentional.

节奏的重点不是停止迭代和改进。

The point of pacing isn't to stop iteration and improvement.

目标是防止更大模型的开发失控。

The goal is to prevent the development of bigger models from spiraling out of control.

Opus和Sol级别的模型现在是我们关注的重点。

Opus and Sol class models are a great place for our focus to go right now.

有很多机会取得真正的成功，而风险没那么大。

Lots of opportunity for real wins without as much risk."

我觉得这个说法是对的。

And I think that that's correct.

如果这就是“控制前沿”时代带来的结果，也就是说，模型能解决之前版本模型的具体问题，在性能和价值上带来渐进式提升，同时在成本和效率上有显著改善，那这似乎是一个值得投入的阶段，而且不会让各种应用的采用速度放慢。

If this is what we get from the pacing the frontier era, in other words, models that solve specific problems with previous iterations of those models and deliver incremental performance or value gains at significant cost and efficiency gains, that seems like a pretty good place to spend some time, and not one where somehow all sorts of adoption is going to slow down.

Claude Code团队的Tarik显然在指出这是实验室的关键方向。

Tarik from the Claude Code team is clearly pointing at this as a key direction for the labs.

他写道：“正确利用模型能力的方式，不是向生产环境发布10倍多的功能。”

He wrote, "The right way to use model capabilities is not to ship 10X more features to prod.

而是花更多时间了解用户，尝试实验，构建原型，学习你不懂的东西，这样才能发布真正有效的产品。”

It's to spend more time understanding your users, trying experiments, building prototypes, learning about things you don't understand so that you can ship things that actually work."

其实，我觉得可以说，因为生成式AI在ChatGPT出现后的整个历史中，能力提升的速度非常快，实验室们可能从来没有真正需要从产品的角度去思考问题。

In fact, I think you could argue that because there have been such radical capability increases at such a fast clip for the entire history of generative AI post-ChatGPT, the labs have never really had to think in product terms.

整个时期里，原始能力的跃升太明显了，他们只需要把最新的东西扔给我们，我们就会照单全收。

The raw capabilities jumps have been so pronounced throughout the entire period that they can just splatter the latest thing at us, and we're going to eat it up.

但到了某个点，愿意这么做的市场就饱和了，那种方式能覆盖的商业用例比例也饱和了，这时候你就得真正考虑人和系统的现实情况了。

But at some point, the market of people who are willing to do that saturates, the percentage of business use cases that that approach can get to saturates, and you actually have to think in terms of human and system realities.

当然，我完全不认为这意味着所有资源会突然都用来完善现有类别的模型，而不是去尝试构建更大的东西，但这确实表明，无休止地追求更大的模型并不是实验室的唯一目标。

Now, I don't at all think that somehow this means that all of a sudden, all the resources are going to go to perfecting existing classes of models rather than trying to build bigger things, but it certainly shows that the endless pursuit of bigger models is not the only be all and end all for the labs.

好了，各位，今天真是相当精彩的一天。

Anyways, guys, it is a pretty phenomenal day.

如果你还没试过，我强烈建议你去花点时间体验一下这些新模型。

If you haven't yet, I would highly encourage you to go out and spend some time with these new models.

那么，本期《AI每日简报》就到这里了。

For now, that is going to do it for this edition of the AI Daily Brief.

一如既往，感谢你的收听或观看。

Appreciate you listening or watching, as always.

下次再见，祝一切顺利。

Until next time, peace.