推出 Claude Opus 5.5

我们推出了 Claude Opus 5.5，这是全新 Claude 5.5 系列中的首个模型。在大多数任务上的表现可媲美 Claude Fable 5.1，而运行成本比 Opus 5 低 40%。We’re introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5.

Claude Opus 5.5 是我们呼吁放缓前沿发展以来的首次发布。该模型在发布前接受了包括 Frontier Design 和 METR 在内的外部评估机构的测试。在我们开展的自动化行为审计中——这是我们运行的最全面的对齐测试——Opus 5.5 是迄今为止我们测试过的表现最强的模型。它还配备了我们为能力最强的模型开发的安全防护措施。Claude Opus 5.5 is our first release since we called for pacing the frontier. It was tested before release by external evaluators, including Frontier Design and METR. On our automated behavioral audit, the most comprehensive alignment test we run, Opus 5.5 is the strongest-performing model we’ve tested to date. It also comes with the safeguards we’ve developed for our most capable models.

以下是您可以期待在 Opus 5.5 中看到的一些改进：Here are some of the improvements you can expect from Opus 5.5:

性能。Opus 5.5 相比 Opus 5 实现了重大提升。它是目前新的领先模型，早期测试者在最复杂的工作中发现其性能大幅跃升。一位测试者在不到一天的时间里完成了一个包含 68 万行代码的迁移——这项工作原本需要一个工程团队耗时数周。它擅长发现并修复软件中的低效问题：当我们要求它缩短 Web 应用每个页面的加载时间时，Opus 5.5 有 40 次中的 39 次成功完成，而 Opus 5 虽然也实现了较小幅度的改进，却同时改变了应用的行为。另一位测试者让多个 Claude 模型仅凭一条提示构建一个游戏；Opus 5.5 的图形表现和精致程度均胜过其他模型，得分最高。Performance. Opus 5.5 is a major step up from Opus 5. It’s the new leading model, and early testers saw large jumps in performance on their most complex work. One tester completed a 680,000-line code migration in less than a day—work that would have taken an engineering team weeks. It’s good at finding and fixing inefficiencies in software: when we asked it to cut load times across every page of a web app, Opus 5.5 succeeded 39 of 40 times, while Opus 5 made smaller improvements that also altered the app’s behavior. A different tester had several Claude models build a game from a single prompt; Opus 5.5 scored higher than any other model on the strength of its graphics and polish.

安全性。Opus 5.5 在我们截至目前的自动化行为审计中取得了所有模型中最优的成绩。这套对齐测试通过数千种模拟场景对 Claude 进行评估。与近期模型相比，它采取难以逆转的行动或越过既定边界行事的可能性大幅降低；与 Opus 5 相比，它也更能抵御提示注入。我们还扩大了对齐测试的范围，涵盖更长的任务、无法完成的任务，以及以真实事件为模型构建的场景，但它仍存在局限。有关评估的完整详情，请参阅《Opus 5.5 系统卡》。Safety. Opus 5.5 achieves the best scores of any model to date on our automated behavioral audit, our alignment suite that tests Claude across thousands of simulated scenarios. It is much less likely than recent models to take hard-to-reverse actions or act outside the boundaries it’s been given, and it’s more resistant than Opus 5 to prompt injection. We’ve also broadened our alignment testing to cover longer tasks, impossible tasks, and scenarios modeled on real incidents, though it still has limits. Full details of our evaluation are available in the Opus 5.5 System Card.

由于 Opus 5.5 在生物学和网络安全方面可与 Claude Mythos 5.1 相媲美，我们将为其部署类似于 Claude Fable 5.1 所采用的安全防护措施。经过审核的组织现可申请加入我们的生命科学验证计划，以使用 Opus 5.5 开展生物学研究。在接下来的几周内，我们还将扩大网络安全验证计划的访问范围，经过验证的网络安全从业者将能够使用 Opus 5.5 开展工作。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. Vetted organizations can apply today to our Life Sciences Verification Program to use Opus 5.5 for biology research. In the coming weeks we will also be expanding access to our Cyber Verification Program, and verified cybersecurity practitioners will be able to use Opus 5.5 for their work.

成本与速度。Opus 5.5 的服务所需计算资源少于 Opus 5，其定价也体现了这一点。我们的测试显示，在默认设置下，对于典型工作负载，其成本将比 Opus 5 低 40%。输入和输出 token 的价格分别为每百万个 4 美元和 20 美元，比 Opus 5 低 20%。缓存读取（占代理和编码工作成本的大多数）为每百万个 token 0.20 美元，比 Opus 5 低 60%。此外，Opus 5.5 生成输出的速度也比 Opus 5 快 30% 以上。Cost and speed. Opus 5.5 requires less compute to serve than Opus 5, and its pricing reflects that. Our tests show that at default settings it will cost 40% less than Opus 5 on typical workloads. Input and output tokens are $4 and $20 per million, 20% less than Opus 5. Cache reads (which make up the majority of agentic and coding work costs) are $0.20 per million tokens, 60% less than Opus 5. Opus 5.5 also generates output more than 30% faster than Opus 5.

除了降价之外，我们还提高了 Pro、Max、Team 和按席位计费的 Enterprise 计划的五小时使用限额。我们还为订阅用户提供了速率限制重置功能，现在你可以保存该重置，并在需要时随时使用。In addition to the price drop, we’re increasing five-hour usage limits on Pro, Max, Team, and seat-based Enterprise plans. We’re also providing subscription users a rate limit reset, which you can now save and use whenever you choose.

沟通。Opus 5.5 比之前的模型沟通得更自然。早期测试者发现，它的文字表达更清晰、更易于理解，这回应了我们听到的关于 Opus 5 的一些常见反馈。它会将最重要的信息放在前面，而其风格也使它在长时间的工作过程中成为更好的合作伙伴。正如一位早期测试者所说：“它的写作方式和我一样。”根据我们自己的使用体验，Opus 5.5 的工作成果更易于理解和检查——这不仅具有实际益处，也有助于提升安全性。Communication. Opus 5.5 communicates more naturally than prior models. Early testers found its writing clearer and easier to follow, which addresses some of the common feedback we heard about Opus 5. It puts the most important information up front, and its style makes it a better work partner over long sessions. As one early tester put it, “it writes the way I do.” In our own use, this has made Opus 5.5’s work easier to follow and check—which is a safety benefit as well as a practical one.

Claude Sonnet 5.5 和 Claude Haiku 5.5 将在未来几周内推出，并带来许多相同的性能、效率和安全性改进。Claude Sonnet 5.5 and Claude Haiku 5.5 will follow in the coming weeks, with many of the same improvements to performance, efficiency, and safety.

性能与成本效益Performance and cost-effectiveness

在我们的基准测试中，Claude Opus 5.5 在代理式编程、计算机使用和知识工作方面处于领先地位。不过，我们发现，在这样的能力水平下，基准测试中的差距已不再是衡量现实世界差异的可靠依据。在我们自己的使用中，Opus 5.5 与 Claude Fable 5.1 之间的差距比这些分数所显示的更小。On our benchmarks, Claude Opus 5.5 leads in agentic coding, computer use, and knowledge work. That said, at these levels of capability we’ve found that benchmark margins have become a less reliable guide to real-world differences. In our own use, the gap between Opus 5.5 and Claude Fable 5.1 is narrower than these scores suggest.

Opus 5.5 的优势在效率方面非常明显。它的每个 token 成本低于 Opus 5，并且每项任务使用的 token 更少，综合下来成本降低了 40%。Where Opus 5.5’s advantage is very clear is efficiency. It costs less per token than Opus 5 and uses fewer tokens per task, which nets out to a 40% drop in costs.

Opus 5.5 的快速模式也已在 Claude Code 和 Claude Platform 中提供，速度最高可提升至 2.5 倍。价格为每百万输入令牌 8 美元，每百万输出令牌 40 美元。Fast mode for Opus 5.5 is also available in Claude Code and the Claude Platform with up to 2.5x speed. It costs $8 per million input tokens and $40 per million output tokens.

编码Coding

Opus 5.5 尤其擅长处理大型且复杂的任务，例如覆盖整个代码库的迁移和审计。一位早期测试者使用它在不到三小时内审计并修复了一个包含 20 万行代码的代码库，而 Opus 5 完成同样任务耗时超过 20 小时，并使用了 2.5 倍的令牌。在一项内部测试中，我们让 Opus 5.5 和 Fable 5.1 将广泛使用的 HAProxy（一款在服务器之间平衡 Web 流量负载的软件）从 C 重写为 Rust。两次重写都通过了 HAProxy 自身几乎所有的回归测试，但 Opus 5.5 用时 9.5 小时完成，而 Fable 5.1 用时 12 小时，成本低了 51%。Opus 5.5 is particularly good at long and sprawling jobs like codebase-wide migrations and audits. An early tester used it to audit and fix a 200,000-line codebase in under three hours, where Opus 5 took over 20 hours and used 2.5x as many tokens. In an internal test, we asked Opus 5.5 and Fable 5.1 to translate HAProxy, widely used software that balances web traffic loads across servers, from C into Rust. Both rewrites passed nearly all of HAProxy’s own regression tests, but Opus 5.5 finished in 9.5 hours compared to 12 for Fable 5.1, and cost 51% less.

Opus 5.5 以极低的成本在智能体编程方面取得了前沿水平的成果。在 FrontierCode 上，按照默认工作强度运行时，它以每项任务约 20% 的成本击败了 GPT-6 Astra。在 Terminal Bench 4.0 上，它以约 40% 的成本达到与 Astra 相当的表现；而在 CursorBench 上，它以约三分之一的成本领先 GPT-5.6 Sol 11 分。Opus 5.5 delivers frontier results on agentic coding at a fraction of the cost. At its default effort level on FrontierCode, it beats GPT-6 Astra at roughly 20% of the cost per task. On Terminal Bench 4.0, it matches Astra for about 40% of the cost, while on CursorBench it beats GPT-5.6 Sol by 11 points for about a third of the cost.

我们的早期测试者报告了类似的效率提升和智能增强：

Our early testers reported similar efficiency and intelligence gains:

最安全的编码代理The most secure coding agent

在系统中使用智能体的企业需要了解这些智能体是否按预期运行，尤其是在它们自主运行数小时的情况下。Opus 5.5 配备了一个在每项操作执行前进行筛查的分类器、一个安全团队可以审计的开源沙箱，以及能够在代码合并前发现漏洞的代码审查机制。Enterprises that use agents within their systems need to know that those agents are operating as intended, particularly when they run autonomously for many hours. Opus 5.5 has a classifier that screens every action before it runs, an open-source sandbox that security teams can audit, and code review that catches vulnerabilities before they merge.

该模型本身也具备更强的防御能力。在提示词注入攻击方面，在我们测试的所有场景中，它的表现都达到或超过了 Opus 5，包括编程、工具使用、计算机操作和网页浏览。在 AI 安全公司 Gray Swan 开展的一项基准测试中，Opus 5.5 与 Fable 5.1 并列成为所有受测模型中提示词注入成功率最低的模型。The model itself also has stronger defenses. On prompt injection attacks, it matches or beats Opus 5 in every setting we tested, including coding, tool use, computer use, and web browsing. On a benchmark run by the AI security firm Gray Swan, Opus 5.5 ties Fable 5.1 for the lowest prompt injection success rate of any model tested.

知识工作Knowledge work

Opus 5.5 是一名可靠且能力出色的研究员。在一项内部测试中，我们要求 Opus 5.5、Fable 5.1 和 Opus 5 仅使用它们在一个难以找到该公司业绩公告的网页副本上找到的信息，撰写一份关于某公司季度业绩的报告。自动评分器将每个数字和引述逐一与来源进行核对。在不同的工作强度设置下，Opus 5.5 撰写的报告中有 18 份中的 16 份达到了我们的质量标准；任何虚构的数字或引述都会导致报告不合格。Fable 5.1 和 Opus 5 在任何一次尝试中都未达到该标准。Opus 5.5 is a reliable and adept researcher. In one internal test, we asked Opus 5.5, Fable 5.1, and Opus 5 to write a report on a company’s quarterly performance using only the information it could find on a copy of the web where the earnings release was hard to locate. An automated grader checked every figure and quote against sources. Across different effort settings, 16 out of 18 of Opus 5.5’s reports cleared our quality bar, where any invented figure or quote would have failed. Neither Fable 5.1 nor Opus 5 cleared that bar in any attempt.

它在金融分析和商业工作方面也表现出色。投资公司兼早期测试者 Walleye Capital 报告称，Opus 5.5 在最低设置下基本解决了他们的评估套件；在更高设置下，表现甚至更好，还发现了其评估说明中的一个错误并对此进行了修正。在此之前，没有其他模型发现过这个错误。It’s also strong in financial analysis and business work. Walleye Capital, an investment firm and early tester, reported that Opus 5.5 largely solved their evaluation suite on its lowest setting; on higher settings, it performed even better, noticing an error in their evaluation instructions and correcting for it. No other model had caught this error before.

在另一项测试中，我们让 Opus 5.5 和 Opus 5 分析两家虚构的人力资源软件公司拟议中的合并。两者都在 Excel 中建立了财务模型，随后将其制作成高管演示文稿，说明按该交易价格进行交易是否合理。两个模型都得出了相同的交易结论，但 Opus 5.5 的模型更为全面，演示文稿也更易于阅读，而 Opus 5 的模型和演示文稿存在一些小错误。Opus 5.5 用时 63 分钟完成，而 Opus 5 用时 93 分钟，且制作成本低 50%。In another test, we tasked both Opus 5.5 and Opus 5 with analyzing a proposed merger between two fictional HR software companies. Each built a financial model in Excel, then turned it into an executive presentation on whether the deal made sense at its price. Both models reached the same conclusions about the deal, but Opus 5.5’s model was more thorough and its presentation easier to read, while Opus 5’s had minor errors. Opus 5.5 finished in 63 minutes compared to 93 for Opus 5, and cost 50% less to produce.

在知识工作评测中，Opus 5.5 的表现优于其他模型，同时使用的令牌更少。在 GDPval-AA v2.1（涵盖 44 个职业的真实工作测试）中，Opus 5.5 获得了 1846 Elo 分，领先于 Fable 5.1 和 Opus 5。在默认工作强度（中等）下，Opus 5.5 以约五分之一的单任务成本，击败了处于最高工作强度的 GPT-6 Astra。在衡量业务工作流和大规模数据收集能力的基准测试中，它同样优于其他模型。On knowledge work evaluations, Opus 5.5 outperforms other models while also using fewer tokens. On GDPval-AA v2.1, a test of real-world work across 44 occupations, Opus 5.5 scores 1846 Elo, ahead of Fable 5.1 and Opus 5. At default effort (medium), Opus 5.5 beats GPT-6 Astra at max effort for about a fifth of the cost per task. It likewise outperformed other models on benchmarks measuring business workflows and large-scale data collection.

我们的客户也反馈了类似的结果。以下是他们告诉我们的与该模型合作的情况：

Our customers have reported similar results. Here’s what they told us about working with the model:

沟通Communication

我们大幅改进了 Opus 5.5 的写作和沟通方式，这是我们听到的关于 Opus 5 最常见的反馈领域之一。它的消息更易于一眼理解，测试人员表示，这在长时间工作期间很有帮助。它会将最重要的信息放在开头，更少使用术语或特立独行的措辞，并遵循你提供的写作规则。我们发现，这使 Opus 5.5 成为了明显更出色的协作者。以下是两个模型的并排比较：We’ve made major improvements to the way Opus 5.5 writes and communicates, one of the most common areas of feedback we heard about Opus 5. Its messages are much easier to understand at a glance, which testers said helped during long working sessions. It puts the most important information up front, is less likely to use jargon or idiosyncratic phrases, and follows the writing rules you give it. We find that this makes Opus 5.5 a noticeably better collaborator. Here’s a side-by-side comparison of the two models:

start, end = month_start(year, month), next_month_start(year, month)
if start <= e.ts < end:

if period.contains(e.ts)

@property
def last_day(self) -> datetime:
    return next_month_start(...) - timedelta(days=1)   # 2026-08-31 00:00:00 UTC

def contains(self, ts):
    return self.first_day <= ts <= self.last_day

start <= e.ts < next_month_start(year, month)    # [Aug 1 00:00, Sep 1 00:00)

last_day = next_month_start(...) - timedelta(days=1)    # = 2026-08-31T00:00:00Z
return self.first_day <= ts <= self.last_day

我们客户的反馈支持这些发现：
Our customers’ feedback supports these findings:

安全Safety

推进前沿Pacing the frontier

上周，我们的首席执行官达里奥·阿莫代伊提出，人工智能的发展速度应当适当控制，以确保安全实践始终领先于模型能力。适度控制发展速度是保障人工智能安全、保持对华竞争力并实现人工智能益处的一种方法，尤其是在生物学和医学等领域。Last week, our CEO, Dario Amodei, argued that AI progress should be paced so that safety practices stay ahead of model capabilities. Pacing is an approach to keeping AI safe, remaining competitive with China, and realizing AI’s benefits, particularly in areas like biology and medicine.

我们基本了解当今模型带来的风险，也有充分能力加以管理。然而，随着模型能力提升，更严重的风险可能会迅速出现，因此我们需要现在就为应对这些风险做好准备。正因如此，我们的安全工作同时着眼于两个时间跨度：We largely understand the risks today’s models present and are well equipped to manage them. However, more serious risks could emerge quickly as capabilities improve, and we need to prepare for them now. For that reason, our safety work takes place on two time horizons at once:

当前模型的安全实践。当前这一代模型依赖一套已确立的实践：广泛的对齐测试、由 METR 和 Frontier Design 等外部组织开展的发布前评估，以及针对每个模型在网络安全和生物学等高风险领域的能力所匹配的防护措施。我们会在每次发布时不断完善这些实践。我们认为，这些实践足以应对当今模型所带来的最严重风险，并且能够让我们大致了解各种严重风险的范围，尽管这种了解并不完美。Safety practices for current models. The current generation of models relies on an established set of practices: extensive alignment testing, pre-release evaluation by outside organizations such as METR and Frontier Design, and safeguards matched to each model’s capabilities in high-risk areas like cybersecurity and biology. We refine these practices with each release. We believe they are appropriate to the worst risks today’s models present, and that they give us a broad, though not perfect, picture of the range of serious risks.

此外，我们还会跟踪训练和评估对齐模型的能力，并在根据《负责任扩展政策》（这是我们用于管理先进人工智能系统灾难性风险的自愿性框架）发布的风险报告中，同时报告我们的公开模型和内部模型。Additionally, we track our ability to train and evaluate aligned models, and we report on both our public and internal models in the risk reports we publish under our Responsible Scaling Policy, our voluntary framework for managing catastrophic risks from advanced AI systems.

为未来的模型做好准备。我们正在为更先进的模型做准备，提前完善训练和评估流程。我们正在加强对强化学习所用环境的筛选，因为存在缺陷的环境是导致行为失准的主要来源之一。此外，我们正在改进对齐奖励，并开发自动化流程，以生成用于安全训练的全新、多样化场景。我们也在加强安全保障和监控工作，包括专注于改进基于可解释性的监控与评估。我们希望，这些技术能够帮助我们减少对审查模型思维链（即模型工作时写出的推理过程）的依赖。Preparing for future models. We’re preparing our training and evaluation processes in anticipation of more advanced models. We’re tightening how we filter the environments used in reinforcement learning, since flawed environments are a major source of misaligned behavior. Additionally, we’re improving our alignment rewards and developing automated processes for producing new, diverse scenarios for safety training. And we are strengthening our security and monitoring, including a focused effort to improve interpretability-based monitoring and evaluation. We hope such techniques will help reduce our reliance on auditing a model’s chain-of-thought, or the reasoning it writes out while it works.

能力更强的模型——例如能够完全自动化 AI 研究本身的模型——需要更高的安全标准。我们呼吁放缓发展步伐，很大程度上是基于这样的预期：此类模型可能很快就会被训练出来。对于这些模型，我们并不认为，仅凭上述措施就能达到这一安全标准。随着 AI 能力不断增强，公共政策应在确保人们所依赖的系统安全方面发挥更大作用。建立这种能力需要时间；正如《我们必须放缓前沿发展》以及我们近期与埃森哲联合发布的公告中所述，我们已经开始建设支持这一目标的基础设施，并预计很快会分享更多相关工作的细节。我们还将继续参与与政府和业界的政策讨论，包括监管方式和国际协调等议题。Models with greater capabilities—such as those that can fully automate the work of AI research itself—require a higher safety standard still. Our calls for pacing were based in large part on our expectation that such models could be trained soon. For these models, we do not assume the measures described above will meet that safety standard on their own. As AI becomes more capable, public policy should play a larger role in making sure the systems people rely on are safe. That capacity takes time to build, and we’ve started to put the infrastructure in place to support it, as described in “We Must Pace the Frontier” and our recent announcement with Accenture; we expect to share more details on these efforts soon. We will also continue to contribute to policy discussions with government and industry, including on approaches to regulation and international coordination.

对齐Alignment

在我们的主要评估套件中，这是一项通过近 2,000 个场景对 Claude 进行评估的自动化行为审计；Opus 5.5 在几乎所有不一致行为指标上的得分都优于近期的任何 Claude 模型。在大多数诚实性指标上，它也是我们表现最强的模型。On our primary evaluation suite, an automated behavioral audit that assesses Claude across nearly 2,000 scenarios, Opus 5.5 scored better than any recent Claude model on nearly every measure of misaligned behavior. It’s also our strongest model on most measures of honesty.

尤其是，Opus 5.5 在多项曾导致近期网络安全事件的行为上优于此前的模型，包括带有偏见或动机驱动的推理、试图逃逸沙箱，以及在得出自己处于模拟环境中的结论后采取有害行动。在一项旨在测试模型跨越遏制边界倾向性的新评估中，Opus 5.5 试图绕过边界的频率比 Opus 5 或 Claude Mythos 5.1 低约 85%，且它进行的每次尝试严重程度都较低，并会主动报告。对于在整个代码库和系统中无人值守地运行 Claude 的团队而言，这一点与原始能力同等重要。In particular, Opus 5.5 improves over previous models on several of the behaviors that contributed to recent cybersecurity incidents, including biased or motivated reasoning, attempting to escape a sandbox, and taking harmful actions after concluding it was in a simulated environment. In a new evaluation designed to test a model’s propensity to cross containment boundaries, Opus 5.5 attempted to circumvent boundaries around 85% less often than Opus 5 or Claude Mythos 5.1, and every attempt it made was low severity and self-reported. For teams running Claude unattended across their codebases and systems, this is just as important as raw capability.

然而，正如我们在近期的对齐评估中所述，构建能够在部署前可靠地捕捉每一种故障的评估仍然是一个尚未解决的问题。我们发现了一些迹象，表明 Opus 5.5 经常会怀疑自己正在接受评估，这使我们难以判断它在所部署的各种现实环境中将如何行动。随着这些环境不断扩展、模型能力不断增强，我们预计这一挑战将日益加剧，除非我们在可解释性方面取得进展。尽管我们确信 Opus 5.5 在我们能够衡量的领域展现出了广泛的改进，但我们仍将自身的对齐工作与下述安全措施结合起来。However, as we described in our recent alignment assessment, building evaluations that reliably catch every failure prior to deployment remains an unsolved problem. We see signs that Opus 5.5 often suspects it is being evaluated, which challenges our ability to assess how it will act in the vast variety of real-world settings it is deployed in. As these settings expand and model capabilities increase, we expect this challenge to grow, unless we make progress on interpretability. Although we are confident that Opus 5.5 shows broad improvements in the areas we are able to measure, we pair our own alignment work with the safeguards described below.

安全措施Safeguards

随着我们的模型变得更加强大，更严格的安全防护措施是我们防止新能力沦为滥用工具的一种方式。Opus 5.5 是首个在网络安全、生物学和蒸馏方面推出与 Fable 5.1 类似级别安全防护措施的 Opus 模型；这些措施都会透明地回退到另一个模型。As our models grow more powerful, stricter safeguards are one way we prevent new capabilities from becoming tools for misuse. Opus 5.5 is the first Opus model to launch with a similar class of safeguards to Fable 5.1 on cybersecurity, biology, and distillation, all of which fall back to another model transparently.

网络安全。由于 Opus 5.5 具备极强的网络安全能力，我们将为 Opus 5.5 采用与 Fable 5.1 类似的网络安全防护措施。用户将能够在常规软件开发生命周期中识别并修复代码中的错误，但大多数网络安全任务将重新转交给 Opus 4.8。Cybersecurity. Because Opus 5.5 has extremely strong cyber capabilities, we’re applying cybersecurity safeguards to Opus 5.5 that are similar to Fable 5.1’s. Users will be able to identify and fix bugs in their code as part of the routine software development lifecycle, but most cybersecurity tasks will be re-routed to Opus 4.8.

对于网络防御人员，我们很快将扩展我们的网络验证计划，纳入 Opus 5.5。新计划将包括三个层级，提供逐步放宽的可信访问权限，其中包括访问 Claude Mythos 模型的权限。Claude Security 已经提供访问 Claude Mythos 5.1 的权限。For cyberdefenders, we’ll soon be expanding our Cyber Verification Program to include Opus 5.5. The new program will include three tiers for increasingly permissive trusted access, including access to Claude Mythos models. Claude Security is already available with access to Claude Mythos 5.1.

生物学。Opus 5.5 在生物学领域能力出众，超越了 Opus 5，并在许多工作领域达到或超过了 Claude Mythos 5.1。例如，在与 Dyno Therapeutics 合作开展的长周期分子预测与设计评估中，Opus 5.5 取得了改进；专家红队测试人员认为，其科学新颖性可与他们测试过的最佳模型相媲美。Biology. Opus 5.5 is highly capable in biology, exceeding Opus 5 and matching or beating Claude Mythos 5.1 across many areas of work. For example, Opus 5.5 achieved improvements on a long-horizon molecular prediction and design evaluation conducted in collaboration with Dyno Therapeutics, and expert red-teamers rated its scientific novelty as comparable to the best model they had tested.

因此，Opus 5.5 采用了与 Fable 5.1 相同的生物学安全防护措施。若要将 Opus 5.5 用于受到这些防护措施限制的研究与开发工作，用户可以申请加入我们新的生命科学验证计划。该计划将为学术实验室、初创企业和制药公司等经过审核的组织提供访问权限，使其能够使用针对广泛生物学相关工作的安全防护措施。有意向的组织可在此申请。For this reason, Opus 5.5 uses the same biology safeguards as Fable 5.1. To use Opus 5.5 for research and development work impeded by these safeguards, users can apply to our new Life Sciences Verification Program, which gives vetted organizations like academic labs, startups, and pharmaceutical companies access to safeguards designed for the full breadth of biology-related work. Interested organizations can apply here.

蒸馏Distillation

蒸馏攻击是指攻击者利用数千个虚假账户，以工业化规模提取模型能力，这会带来安全和国家安全风险。蒸馏使不法行为者能够创建高度强大的模型，却不具备我们融入 Claude 的安全防护措施。我们 2026 年 9 月的威胁情报报告详细介绍了迄今为止我们发现并遏制的非法蒸馏活动。Distillation attacks, in which attackers use thousands of fake accounts to extract a model’s capabilities at industrial scale, create safety and national security risks. Distillation allows bad actors to create highly capable models without the safeguards we build into Claude. Our September 2026 threat intelligence report details the illicit distillation activity we’ve detected and disrupted so far.

Opus 5.5 发布时将启用保留思考功能，这是我们在 Fable 5.1 中引入的反蒸馏保护措施。它可以阻止 API 用户编辑 Claude 之前的上下文，以试图提取 Claude 的推理过程。该措施适用于 2026 年 8 月 31 日或之后创建的 API 账户上的 Fable 5.1 和 Opus 5.5。我们的帮助中心文章解释了这一变更，而保留思考文档则介绍了如何测试并更新您的集成。Opus 5.5 is launching with preserved thinking, the anti-distillation safeguard we introduced with Fable 5.1. It stops API users from editing Claude’s prior context in an attempt to extract Claude’s reasoning. It applies to Fable 5.1 and Opus 5.5 for API accounts created on or after August 31, 2026. Our Help Center article explains the change, and our preserved thinking docs show how to test and update your integrations.

数据保留与合规Data retention and compliance

与之前的 Opus 模型一样，Opus 5.5 支持零数据保留。Like previous Opus models, Opus 5.5 is available with zero data retention.

与 Fable 5.1 一样，Opus 5.5 也采用了我们的水印措施，以遵守《欧盟人工智能法案》，相关内容请参阅此处。此外，Opus 5.5 也不再提供关闭“思考”模式的选项，具体说明请参阅此处。As with Fable 5.1, Opus 5.5 comes with our watermarking measures to comply with the EU AI Act, discussed here. It is also no longer available with “thinking” mode switched off, as we describe here.

可用性Availability

Claude Opus 5.5 is now available on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure. On the Claude Platform, developers can get started with claude-opus-5-5.