Product产品Announcements公告

Introducing Claude Opus 5隆重推出 Claude Opus 5

2026年7月24日2026年7月24日
Introducing Claude Opus 5

Claude Opus 5 is available today. It’s a thoughtful and proactive model that comes close to the frontier intelligence of Claude Fable 5 at half the price.Claude Opus 5 即日起正式发布。这是一款深思熟虑且具有主动性的模型,其智能水平已接近 Claude Fable 5 的前沿水准,而价格仅为其一半。

On coding and knowledge work evaluations like Frontier-Bench and GDPval-AA, Opus 5 is the new state-of-the-art, though it remains behind Mythos 5 on cybersecurity tasks.在 Frontier-Bench 和 GDPval-AA 等编程与知识工作评估中,Opus 5 树立了新的行业标杆,不过在网络安全任务方面仍落后于 Mythos 5。

Opus 5 is designed to be used every day: it works more efficiently than other models. It’s the new default model on Claude Max, and the strongest model on Claude Pro.Opus 5 专为日常使用而设计,运行效率高于其他模型。它是 Claude Max 的默认模型,也是 Claude Pro 中性能最强的模型。

Performance and cost-effectiveness性能与成本效益

Claude Opus 5 provides greatly improved performance for the same cost as its predecessor, Opus 4.8. The charts in this section show how performance changes according to the model’s effort setting, which customers can use to optimize for intelligence or conserve tokens for faster and cheaper results.Claude Opus 5 在保持与前代产品 Opus 4.8 相同成本的同时,实现了性能的显著提升。本节图表展示了性能如何随模型的“努力程度”(effort)设置而变化,客户可以利用此设置来优化智能水平,或通过节省 Token 来获得更快、更经济的结果。

Opus 5 excels on valuable software engineering tasks. For example, on Frontier-Bench v0.1, Opus 5 surpasses all other models, and more than doubles Opus 4.8’s performance at a lower cost per task. On CursorBench 3.2, at max effort, the model performs within 0.5% of Fable 5’s peak score, but at half the cost per task; it also achieves greater performance at a given cost than all other models on high, xhigh, and max effort.Opus 5 在高价值软件工程任务中表现卓越。例如,在 Frontier-Bench v0.1 上,Opus 5 超越了所有其他模型,且以更低的单任务成本实现了两倍于 Opus 4.8 的性能。在 CursorBench 3.2 上,当设置为最大努力程度时,该模型的性能与 Fable 5 的峰值得分相差不到 0.5%,但单任务成本却降低了一半;此外,在“高”、“超高”和“最大”努力程度下,它在给定成本下的性能表现均优于所有其他模型。

We see similar results on knowledge work and problem-solving tasks. For example:我们在知识工作和问题解决任务中也看到了类似的结果。例如:

  • On ARC-AGI 3, an evaluation where the model has to solve novel problems, Opus 5’s score is three times as high as the next-best model.在 ARC-AGI 3(一项要求模型解决新颖问题的评估)中,Opus 5 的得分是排名第二模型的 3 倍。
  • On Zapier AutomationBench, which measures whether models can complete business tasks from start to finish, Opus 5’s pass rate is around 1.5× the next-best model for the same cost per task. Even at its lowest effort setting, Opus 5 passes more tasks than any other model.在衡量模型能否从头到尾完成业务任务的 Zapier AutomationBench 上,Opus 5 的任务通过率约为排名第二模型的 1.5 倍(成本相同)。即使在最低努力程度设置下,Opus 5 完成的任务量也超过了任何其他模型。
  • On OSWorld 2.0, a computer use benchmark, Opus 5 outperforms every other model at any given cost, surpassing Fable 5’s best result at just over a third of the cost.在计算机操作基准测试 OSWorld 2.0 中,Opus 5 在任何成本下都优于其他所有模型,仅以三分之一多一点的成本就超越了 Fable 5 的最佳成绩。

It’s also our best and most cost-efficient model on several related evaluations:它也是我们在多项相关评估中性能最强且最具成本效益的模型:

Opus 5 is a meaningful improvement over Opus 4.8 for scientific research. It shows better performance than Opus 4.8 on every one of our life sciences evaluations, which cover topics including structural biology, organic chemistry, and bioinformatics. Its improvements are most notable on organic chemistry tasks, like inferring molecular structures from spectroscopy data (it scores 10.2 percentage points higher than Opus 4.8 on our internal benchmark), and on protein-related tasks like predicting how variations in a protein’s sequence affect how it functions (here, it scores 7.7 percentage points higher).对于科学研究而言,Opus 5 相比 Opus 4.8 有了显著进步。在我们涵盖结构生物学、有机化学和生物信息学等主题的所有生命科学评估中,它的表现均优于 Opus 4.8。其改进在有机化学任务中尤为突出,例如从光谱数据推断分子结构(在我们内部基准测试中得分比 Opus 4.8 高出 10.2 个百分点),以及蛋白质相关任务,如预测蛋白质序列变异如何影响其功能(得分高出 7.7 个百分点)。

Finally, Opus 5 is capable of producing much stronger visual outputs:最后,Opus 5 能够生成更出色的视觉输出:

Working with Claude Opus 5与 Claude Opus 5 协作

Claude Opus 5 is much stronger at verifying its work and iterating carefully until it succeeds. In evaluations and early-access testing, we and our users found many examples of Opus 5’s agency and thoroughness:Claude Opus 5 在验证自身工作并仔细迭代直至成功方面能力更强。在评估和抢先体验测试中,我们和用户发现了许多体现 Opus 5 主动性和严谨性的案例:

  • On one Frontier-Bench task, Opus 5 was given a drawing of a machine part and asked to write code to rebuild it as a 3D FreeCAD model. However, in this task, the model was intentionally given no way to directly view the drawing. Opus 5 responded by writing its own computer vision pipeline to pull the geometry from the raw pixels, then reconstructed the full machine part. It succeeded in doing so repeatedly; no competing model with the same setup could solve it after five attempts.在一项 Frontier-Bench 任务中,Opus 5 获得了一张机械零件图纸,并被要求编写代码将其重建为 3D FreeCAD 模型。然而,在这个任务中,模型被刻意限制无法直接查看图纸。Opus 5 的应对方式是编写了自己的计算机视觉流水线,从原始像素中提取几何形状,然后重建了完整的机械零件。它多次成功完成了任务;而竞争模型在相同设置下尝试五次后均无法解决。
  • Given a real bug in a popular open-source package manager, Opus 5 found the root cause and fixed an edge case that the community’s patch had missed. A competing model fixed only the surface symptom (not the underlying cause), then reported the bug resolved.针对一个流行开源包管理器中的真实漏洞,Opus 5 找到了根本原因并修复了一个社区补丁遗漏的边缘情况。而竞争模型仅修复了表面症状(而非根本原因),随后便报告漏洞已解决。
  • An engineer at a trading firm used Opus 5 to build a market data feed for a new exchange in a single session. Previous models could not complete this task at all, even given extensive plans from the engineer. Finding no live feed to validate against, Opus 5 even built its own test harness to check that its code parsed the exchange’s data correctly.一家贸易公司的工程师使用 Opus 5 在单次会话中构建了一个新交易所的市场数据馈送。之前的模型即便在获得工程师详细计划的情况下也无法完成此任务。由于没有实时馈送进行验证,Opus 5 甚至自行构建了一个测试工具来检查其代码是否正确解析了交易所的数据。

Below are further reports from our early-access customers on their experience of working with Opus 5:以下是我们抢先体验客户在使用 Opus 5 后的更多反馈:

Alignment and safety对齐与安全

Alignment. During pre-deployment testing, our automated behavioral audit found Opus 5 to be our most aligned model to date (as shown in the graph below). It adheres to Claude’s Constitution better than Opus 4.8, Sonnet 5, or Fable 5; exhibits the lowest rates of deceptive behavior; and is the least susceptible to being tricked into misuse. It’s also our safest model yet in terms of avoiding reckless actions that could have hard-to-reverse side effects.对齐。在部署前测试期间,我们的自动化行为审计发现 Opus 5 是我们迄今为止最对齐的模型(如下图所示)。它比 Opus 4.8、Sonnet 5 或 Fable 5 更能遵守 Claude 的宪法;表现出最低的欺骗性行为发生率;并且最不容易被诱导进行滥用。就避免可能产生难以逆转的副作用的鲁莽行为而言,它也是我们迄今为止最安全的模型。

On our automated behavioral audit, Opus 5 scores 2.3 on overall misaligned behavior, the lowest of our recent models.在我们的自动化行为审计中,Opus 5 在整体不对齐行为方面的得分为 2.3,是我们近期模型中最低的。

Safety. Opus 5 does not advance the frontier in risky, dual-use capabilities. In rigorous evaluations conducted alongside private-sector and government partners, we found it remains behind Mythos 5 in both biology research and offensive cybersecurity. More information about these evaluations can be found in our System Card.安全。Opus 5 没有在风险性、双重用途能力方面推进前沿。在与私营部门和政府合作伙伴共同进行的严格评估中,我们发现它在生物研究和进攻性网络安全方面仍落后于 Mythos 5。有关这些评估的更多信息,请参阅我们的系统卡。

As with its predecessor, Opus 4.8, we’ve intentionally avoided training Opus 5 on cyber tasks. The model has nevertheless improved substantially on these tasks as a result of becoming more generally capable, and it comes close to Mythos 5 at finding cybersecurity vulnerabilities. However, it remains substantially behind Mythos 5 on the exploitation of those vulnerabilities—that is, in turning vulnerabilities into material cyber threats.与前代产品 Opus 4.8 一样,我们特意避免在网络任务上训练 Opus 5。然而,由于变得更加通用,该模型在这些任务上的表现仍然有了显著提高,在发现网络安全漏洞方面已接近 Mythos 5。不过,在利用这些漏洞方面,它仍显著落后于 Mythos 5——即在将漏洞转化为实质性网络威胁方面。

This is illustrated by Opus 5’s performance on OSS-Fuzz, an evaluation we’ve developed to assess how well models can find and then exploit vulnerabilities without extensive human guidance. Although Mythos 5 and Opus 5 identify vulnerabilities with similar success, Opus 5’s score on the development of exploits is far behind that of Mythos 5.Opus 5 在 OSS-Fuzz 上的表现说明了这一点,这是我们开发的一项评估,旨在评估模型在没有大量人工指导的情况下发现并利用漏洞的能力。尽管 Mythos 5 和 Opus 5 在识别漏洞方面的成功率相似,但 Opus 5 在开发漏洞利用程序方面的得分远落后于 Mythos 5。

On OSS-Fuzz, one of our cybersecurity evaluations, Opus 5 is close to Mythos 5 at identifying software vulnerabilities (left), but is considerably less successful at developing exploits for them (right).在我们的网络安全评估之一 OSS-Fuzz 上,Opus 5 在识别软件漏洞方面(左)接近 Mythos 5,但在开发漏洞利用程序方面(右)却远不如后者成功。

Safeguards for Opus 5Opus 5 的保障措施

Claude Opus 5’s safeguards are designed to allow beneficial uses of the model in both cybersecurity and biology. They are similar to those we applied to Opus 4.8, with the exception of some stronger guardrails on a narrow range of cyber tasks.Claude Opus 5 的保障措施旨在允许在网络安全和生物学领域对模型进行有益的使用。它们类似于我们应用于 Opus 4.8 的措施,只是在极少数网络任务上设置了更强的护栏。

Cybersecurity. Opus 5’s cyber classifiers are proportionally less restrictive than those on Fable 5. They allow Opus 5 to find vulnerabilities in source code, but block “binary-based” vulnerability scanning (a method more likely to be associated with malicious actors), penetration testing, and exploit generation.网络安全。Opus 5 的网络分类器在比例上比 Fable 5 的限制更少。它们允许 Opus 5 在源代码中查找漏洞,但会阻止“基于二进制”的漏洞扫描(一种更容易与恶意行为者关联的方法)、渗透测试和漏洞利用生成。

Based on our testing, we expect the classifiers to intervene around 85% less often than they do for Fable 5. In Claude.ai, Claude Code, and Claude Cowork, any flagged requests will fall back to Opus 4.8 by default. Fallbacks to Opus 4.8 can also be enabled on the API.根据我们的测试,我们预计分类器的干预频率将比 Fable 5 低约 85%。在 Claude.ai、Claude Code 和 Claude Cowork 中,任何被标记的请求将默认回退到 Opus 4.8。回退到 Opus 4.8 也可在 API 上启用。

Our Cyber Verification Program (CVP) facilitates cybersecurity work that would otherwise be impeded by the model’s safeguards. Enterprises and researchers who are already part of the CVP have immediate access to a version of Opus 5 with fewer security restrictions.我们的网络验证计划 (CVP) 促进了网络安全工作,否则这些工作会受到模型保障措施的阻碍。已经是 CVP 一部分的企业和研究人员可以立即访问具有较少安全限制的 Opus 5 版本。

Biology. Since Opus 5 has a similar suite of safeguards to Opus 4.8, it is now our most capable generally available model for scientific research. Nevertheless, the model still shows important limitations on long-running, autonomous research tasks, which is where we expect AI models to pose the most substantial biology-related risks. (Mythos 5 remains the stronger model for this type of biological work.) As part of this launch, biology-related requests that are blocked on Fable 5 will now route to Opus 5 rather than Opus 4.8.生物学。由于 Opus 5 拥有一套与 Opus 4.8 类似的保障措施,它现在是我们用于科学研究的最强大的通用模型。尽管如此,该模型在长期、自主的研究任务上仍然表现出重要的局限性,而这正是我们预计 AI 模型会带来最重大生物相关风险的地方。(Mythos 5 仍然是此类生物学工作的更强模型。)作为此次发布的一部分,在 Fable 5 上被阻止的生物相关请求现在将路由到 Opus 5,而不是 Opus 4.8。

Getting started入门

Claude Opus 5 is available today on all platforms, priced at $5 per million input tokens and $25 per million output tokens (the same as Opus 4.8). Developers can get started with claude-opus-5 on the Claude API.Claude Opus 5 即日起在所有平台上可用,定价为每百万输入 Token 5 美元,每百万输出 Token 25 美元(与 Opus 4.8 相同)。开发者可以在 Claude API 上使用 claude-opus-5 开始使用。

It’s also offered in Fast mode, where it runs around 2.5 times the default speed. As with Opus 4.8, Fast mode is available at twice Opus 5’s base price on the Claude Platform and through usage credits in Claude Code.它还以“快速模式”(Fast mode)提供,运行速度约为默认速度的 2.5 倍。与 Opus 4.8 一样,快速模式在 Claude 平台上以 Opus 5 基本价格的两倍提供,并通过 Claude Code 中的使用额度提供。

Alongside Opus 5, we’re releasing two updates in beta:除了 Opus 5,我们还发布了两个处于测试阶段的更新:

  • Mid-conversation tool changes on the Claude Platform. Within a conversation, developers can now change which tools Claude can use without invalidating the prompt cache.Claude 平台上的对话中途工具更改。在对话中,开发者现在可以更改 Claude 可以使用的工具,而不会使 Prompt 缓存失效。
  • Automatic fallbacks on the API. Users can now choose to have requests that are flagged by our safety classifiers on Opus 5 (or Fable 5) automatically route to another model. With automatic fallbacks on, API requests always route to the best available model by default rather than being blocked.API 上的自动回退。用户现在可以选择让被 Opus 5(或 Fable 5)安全分类器标记的请求自动路由到另一个模型。开启自动回退后,API 请求默认总是路由到最佳可用模型,而不是被阻止。

Consistent with prior Opus models, Opus 5 does not have data retention requirements for general access.与之前的 Opus 模型一致,Opus 5 对于通用访问没有数据保留要求。

For more guidance on how to get the best out of Opus 5, see our prompting guide.有关如何充分利用 Opus 5 的更多指导,请参阅我们的提示词指南。

Footnotes脚注

Frontier-Bench v0.1, Effort plot: These results are from an internal run of Frontier-Bench v0.1, on the mini-SWE-agent harness and a GKE backend, mean reward over 5 attempts per task. Opus 4.8 served as fallback on safety-classifier refusals for Opus 5 and Fable 5.Frontier-Bench v0.1,努力程度图:这些结果来自 Frontier-Bench v0.1 的内部运行,基于 mini-SWE-agent 工具和 GKE 后端,每个任务 5 次尝试的平均奖励。Opus 4.8 用作 Opus 5 和 Fable 5 在安全分类器拒绝时的回退模型。

Related content相关内容

A research agenda for the Economic Futures Research Fund经济未来研究基金的研究议程

We’re sharing the research agenda for the Anthropic Economic Futures Research Fund.我们正在分享 Anthropic 经济未来研究基金的研究议程。

Read more阅读更多

Ask Claude about the Anthropic Economic Index向 Claude 询问 Anthropic 经济指数

We're launching the Anthropic Economic Index connector for Claude, which lets anyone explore real data about AI and work.我们正在为 Claude 推出 Anthropic 经济指数连接器,让任何人都可以探索关于 AI 和工作的真实数据。

Read more

Anthropic is donating another $20 million to Public First ActionAnthropic 向 Public First Action 再捐赠 2000 万美元

Anthropic is contributing an additional $20 million to Public First Action, bringing our total support to $40 million.Anthropic 正在向 Public First Action 再捐赠 2000 万美元,使我们的总支持额达到 4000 万美元。

Read more