Anthropic发布Claude Fable 5.2与Mythos 5.2模型
Claude Fable 5.2 and Mythos 5.2
Anthropic旗舰技术模型迭代，性能全面超越现有Opus系列并压制主要竞品，开发者应关注其新定价策略与Agent能力突破。
We’ve upgraded Claude Fable and Claude Mythos—models that are designed to tackle the most daunting technical challenges.
我们已升级 Claude Fable 和 Claude Mythos——这两款模型旨在应对最艰巨的技术挑战。
Claude Fable 5.2 and Claude Mythos 5.2 show improvements in areas like coding, reasoning, and vision. During prerelease testing they’ve helped our engineers complete otherwise-infeasible software migrations, designed three working electronic devices from scratch, solved two open problems in the physics of black holes, made substantial progress on robotics evaluations, and formalized the proof of Fermat’s Last Theorem in 30 hours—among many other applications we’ll discuss below.
Claude Fable 5.2 和 Claude Mythos 5.2 在编码、推理和视觉等领域展现出改进。在预发布测试期间，它们帮助我们的工程师完成了原本不可行的软件迁移，从零设计了三个可工作的电子设备，解决了黑洞物理学中的两个开放性问题，在机器人评估方面取得了实质性进展，并在 30 小时内形式化了费马大定理的证明——此外还有我们将在下文讨论的许多其他应用。
Claude Fable and Claude Mythos are the same underlying model. Fable has strong cybersecurity and biology safeguards to guard against misuse, which allows us to make it available to all users. Organizations pursuing beneficial applications in cybersecurity or biomedical research can apply for access to Mythos through our recently expanded Life Sciences and Cyber Verification Programs.
Claude Fable 和 Claude Mythos 基于相同的底层模型。Fable 具备强大的网络安全和生物学安全护栏，以防止滥用，这使得我们能够向所有用户开放使用。致力于在网络安全或生物医学研究中进行有益应用的组织，可通过我们近期扩展的生命科学与网络安全验证计划申请访问 Mythos。
Testers have found that Fable 5.2 and Mythos 5.2 communicate more clearly than past Fable and Mythos models. They generate outputs up to 30% faster than Claude Fable 5.1 and Claude Mythos 5.1, and they’re available at a lower price: $8 per million input tokens and $40 per million output tokens.
测试人员发现，Fable 5.2 和 Mythos 5.2 的沟通清晰度优于以往的 Fable 和 Mythos 模型。它们的输出速度比 Claude Fable 5.1 和 Claude Mythos 5.1 快多达 30%，且价格更低：每百万输入令牌 8 美元，每百万输出令牌 40 美元。
Fable 5.2 and Mythos 5.2 also come with an optional fast mode, which speeds up token generation by up to 2.5 times. Fast mode is available for $16 per million input tokens and $80 per million output tokens.
Fable 5.2 和 Mythos 5.2 还附带可选的快速模式，可将令牌生成速度提升高达 2.5 倍。快速模式的定价为每百万输入令牌 16 美元，每百万输出令牌 80 美元。
Safety and alignment. To help ensure safety and alignment keep step with growing model capabilities, we ran extensive prerelease safety testing on the model that underlies Fable 5.2 and Mythos 5.2. This included developing novel evaluations that give deeper insight into specific concerning behaviors like sandbox escapes. We believe high-stakes risks from misalignment remain low, and the model performed as well as, or better than, our recent models on most alignment measures. Fable 5.2’s cyber and biology safeguards are as robust as those of Fable 5.1’s but produce fewer unnecessary blocks, and we have implemented our strongest-ever protections against distillation attacks.
安全与对齐。为确保安全与对齐能够跟上模型能力的不断提升，我们对构成 Fable 5.2 和 Mythos 5.2 的底层模型进行了广泛的预发布安全测试。其中包括开发新型评估方法，以更深入地了解沙箱逃逸等特定令人担忧的行为。我们认为，由不对齐引发的高风险仍然较低，并且该模型在大多数对齐指标上的表现与我们的近期模型相当或更优。Fable 5.2 的网络安全和生物学安全护栏与 Fable 5.1 一样稳健，但产生的误拦截更少，我们还实施了迄今为止最强的防蒸馏攻击保护措施。
We describe more about our safety testing below, and full details can also be found in the Claude Fable 5.2 and Mythos 5.2 system card.
我们在下文将详细描述我们的安全测试情况，完整细节也可在 Claude Fable 5.2 和 Mythos 5.2 系统卡片中找到。
When to use Claude Fable 5.2 and Mythos 5.2. Fable and Mythos are best for the most ambitious, open-ended technical work. For most coding and knowledge tasks, we recommend that users start with Claude Opus 5.5, our strongest daily driver, and then switch to a different model as required, depending on the scope and complexity of their work—upward to Fable or Mythos, or downward to Claude Sonnet 5.5 or Claude Haiku 5.5. Fable 5.2 can also play the role of Advisor (where Haiku, Sonnet, or Opus models consult it to help solve harder problems) or Orchestrator (where it delegates simpler tasks to smaller models).
何时使用 Claude Fable 5.2 和 Mythos 5.2。Fable 和 Mythos 最适合最具雄心、开放式的技术工作。对于大多数编码和知识任务，我们建议用户从 Claude Opus 5.5（我们最强的日常主力模型）开始，然后根据工作的范围和复杂度切换到不同的模型——向上切换至 Fable 或 Mythos，或向下切换至 Claude Sonnet 5.5 或 Claude Haiku 5.5。Fable 5.2 还可以扮演顾问（Advisor）角色（Haiku、Sonnet 或 Opus 模型会向其咨询以协助解决更复杂的问题），或充当编排器（Orchestrator）（将较简单的任务委派给较小的模型）。
Evaluating Claude Fable 5.2 and Mythos 5.2
评估 Claude Fable 5.2 和 Mythos 5.2
Benchmark performance. The benchmarks table below shows that Claude Fable 5.2 and Mythos 5.2 are a step above Opus 5.5 in almost all the capabilities we evaluated:
基准性能。下表中的基准测试结果显示，Claude Fable 5.2 和 Mythos 5.2 在我们评估的大多数能力上均优于 Opus 5.5：
| Fable 5.2 | Fable 5.1 | Opus 5.5 | GPT-6 Astra | Gemini 4 Carbon | |
|---|---|---|---|---|---|
| Agentic codingFrontierCode 2.0 | |||||
| Agentic codingFrontierCode 2.0 | 49.1%Xhigh effort | 32.4%Xhigh effort | 45.5% | 42.8% | 41.1%Medium effort |
| Agentic codingFrontierSWE v2 | |||||
| Agentic codingFrontierSWE v2 | 69.9% | 56.3% | 62.3% | 65.5% | 61.2% |
| Business workflowsAutomationBench | |||||
| Business workflowsAutomationBench | 51.4% | 31.4% | 42.5% | 41.4% | 53.9% |
| Agentic scientific researchTerminal-Bench-Science 0.1 | |||||
| Agentic scientific researchTerminal-Bench-Science 0.1 | 74.6%75.3% (Mythos 5.2) | 52.6% | 58.7% | 68.1% | 64.8% |
| CybersecurityCyScenarioBench† | |||||
| CybersecurityCyScenarioBench† | 76.3%Mythos 5.2 | 61.7%Mythos 5.1 | 67.6%without safeguards | 61%GPT-6 Cyber | — |
| CybersecurityCVE Rediscovery | |||||
| CybersecurityCVE Rediscovery | 89.6%Mythos 5.2 | 78.3%Mythos 5.1 | 81.3%without safeguards | — | — |
| Protein designDe novo binder design (24 hr) | |||||
| Protein designDe novo binder design (24 hr) | 85.1%Mythos 5.2 | 79.1%Mythos 5.1 | 82.6%without safeguards | — | — |
| Agentic searchWANDR (relaxed score) | |||||
| Agentic searchWANDR (relaxed score) | 82.1% | 68.7% | 72.3% | — | — |
| Computer useOSWorld 2.1 (offline set, partial) | |||||
| Computer useOSWorld 2.1 (offline set, partial) | 89.3% | 83.5% | 87.2% | 79.5% | 81.5% |
| RoboticsRoboDojo Sim‡ | |||||
| RoboticsRoboDojo Sim‡ | 64.9% | 27.6% | 39.4% | 29.8% | — |
| Knowledge workAA-Briefcase v1.1 | |||||
| Knowledge workAA-Briefcase v1.1 | 66.1%rubric | 61.5%rubric | 61.2%rubric | 52.0%rubric | 61.8%rubric |
| 1785Elo | 1669Elo | 1800Elo | 1569Elo | 1611Elo | |
| Knowledge workRemote Labor Index (RLI) | |||||
| Knowledge workRemote Labor Index (RLI) | 41.3%High effort | 17.9% | 21.3% | 20.8% | — |
| Multidisciplinary reasoningHumanity’s Last Exam (with tools) | |||||
| Multidisciplinary reasoningHumanity’s Last Exam (with tools) | 74.2% | 65.6% | 67.7% | 57.2% | — |
| Visual chart recognitionChartography (no tools) | |||||
| Visual chart recognitionChartography (no tools) | 73.4% | 44.8% | 64.4% | 71.0% | 72.5% |
| Fable 5.2 | Fable 5.1 | Opus 5.5 | GPT-6 Astra | Gemini 4 Carbon | |
|---|---|---|---|---|---|
| 智能体编码FrontierCode 2.0 | |||||
| 智能体编码FrontierCode 2.0 | 49.1%高投入 | 32.4%高投入 | 45.5% | 42.8% | 41.1%中等投入 |
| 智能体编码FrontierSWE v2 | |||||
| 智能体编码FrontierSWE v2 | 69.9% | 56.3% | 62.3% | 65.5% | 61.2% |
| 业务流程自动化AutomationBench | |||||
| 业务流程自动化AutomationBench | 51.4% | 31.4% | 42.5% | 41.4% | 53.9% |
| 智能体科学研究Terminal-Bench-Science 0.1 | |||||
| 智能体科学研究Terminal-Bench-Science 0.1 | 74.6%75.3% (Mythos 5.2) | 52.6% | 58.7% | 68.1% | 64.8% |
| 网络安全CyScenarioBench† | |||||
| 网络安全CyScenarioBench† | 76.3%Mythos 5.2 | 61.7%Mythos 5.1 | 67.6%无安全限制 | 61%GPT-6 Cyber | — |
| 网络安全CVE Rediscovery | |||||
| 网络安全CVE Rediscovery | 89.6%Mythos 5.2 | 78.3%Mythos 5.1 | 81.3%无安全限制 | — | — |
| 蛋白质设计De novo binder design (24 hr) | |||||
| 蛋白质设计De novo binder design (24 hr) | 85.1%Mythos 5.2 | 79.1%Mythos 5.1 | 82.6%无安全限制 | — | — |
| 智能体搜索WANDR (relaxed score) | |||||
| 智能体搜索WANDR (relaxed score) | 82.1% | 68.7% | 72.3% | — | — |
| 计算机使用OSWorld 2.1 (offline set, partial) | |||||
| 计算机使用OSWorld 2.1 (offline set, partial) | 89.3% | 83.5% | 87.2% | 79.5% | 81.5% |
| 机器人技术RoboDojo Sim‡ | |||||
| 机器人技术RoboDojo Sim‡ | 64.9% | 27.6% | 39.4% | 29.8% | — |
| 知识工作AA-Briefcase v1.1 | |||||
| 知识工作AA-Briefcase v1.1 | 66.1%rubric | 61.5%rubric | 61.2%rubric | 52.0%rubric | 61.8%rubric |
| 1785Elo | 1669Elo | 1800Elo | 1569Elo | 1611Elo | |
| 知识工作Remote Labor Index (RLI) | |||||
| 知识工作Remote Labor Index (RLI) | 41.3%高投入 | 17.9% | 21.3% | 20.8% | — |
| 多学科推理Humanity’s Last Exam (with tools) | |||||
| 多学科推理Humanity’s Last Exam (with tools) | 74.2% | 65.6% | 67.7% | 57.2% | — |
| 图表识别Chartography (no tools) | |||||
| 图表识别Chartography (no tools) | 73.4% | 44.8% | 64.4% | 71.0% | 72.5% |
For details on evaluation methodology, see the Claude Fable 5.2 and Claude Mythos 5.2 System Card. Scores for Fable 5.2, Fable 5.1, and Opus 5.5 show results with safeguards on unless otherwise noted. All scores show each model’s highest-available effort unless otherwise noted. † On CyScenarioBench, each lab’s scores are based on unique tasks. ‡ RoboDojo Sim shows the highest-available effort and was run internally.
有关评估方法的详细信息，请参阅 Claude Fable 5.2 和 Claude Mythos 5.2 系统卡片。除非另有说明，Fable 5.2、Fable 5.1 和 Opus 5.5 的得分均显示启用安全限制时的结果。除非另有说明，所有得分均显示每个模型可用的最高投入水平。† 在 CyScenarioBench 上，各实验室的得分基于独特的任务。‡ RoboDojo Sim 显示的是可用的最高投入水平，并在内部运行。
Effort and efficiency. On Medium effort (the default in Claude apps and Claude Code), Fable 5.2 can accomplish challenging tasks with fewer tokens than Fable 5.1. In early testing on the FrontierCode 2.0 evaluation, Fable 5.2 at High effort scored about level with Opus 5.5 at Max effort, for roughly half the cost per task. At Xhigh effort, it also scored the highest of any released model tested.
努力程度与效率。在中等努力程度（Claude 应用和 Claude Code 中的默认设置）下，Fable 5.2 完成具有挑战性的任务所需的 token 数少于 Fable 5.1。在 FrontierCode 2.0 评估的早期测试中，Fable 5.2 在高努力程度下的得分与 Opus 5.5 在最大努力程度下的得分大致相当，但每项任务的成本约为后者的一半。在极高努力程度下，其得分也是我们所测试的所有已发布模型中最高的。
Agentic coding: FrontierCode 2.0Agentic search: WANDRHardware engineering: BenchCAD 2.1Multidisciplinary reasoning: Humanity’s Last ExamVisual chart recognition: Chartography
智能体编码：FrontierCode 2.0 智能体搜索：WANDR 硬件工程：BenchCAD 2.1 多学科推理：Humanity’s Last Exam 视觉图表识别：Chartography
Agentic coding: FrontierCode 2.0Agentic search: WANDRHardware engineering: BenchCAD 2.1Multidisciplinary reasoning: Humanity’s Last ExamVisual chart recognition: Chartography
智能体编码：FrontierCode 2.0 智能体搜索：WANDR 硬件工程：BenchCAD 2.1 多学科推理：Humanity’s Last Exam 视觉图表识别：Chartography
FrontierCode 2.0Accuracy vs. cost
FrontierCode 2.0 准确率与成本
- Fable 5.2
- Opus 5.5
- Fable 5.1
- GPT-6 Astra
- Gemini 4 Carbon
10203040500Score (%)125102050Cost per task (USD, log scale)LowMedHighXhighMax
10 20 30 40 50 0 分数 (%) 1 2 5 10 20 50 每项任务成本 (美元，对数刻度) 低 中 高 极高 最大
FrontierCode measures whether a model’s code changes would be merged by the maintainers of open-source repositories. At High effort, Fable 5.2 scores about the same as Opus 5.5’s best score (45.3% vs. 45.5%) for roughly half the cost per task, and ahead of GPT-6 Astra’s best (42.8%) at about the same cost. At Xhigh effort, Fable 5.2 reaches 49.1%, the highest score of any released model we’ve tested.
FrontierCode 衡量的是开源仓库维护者是否会合并模型的代码更改。在高努力程度下，Fable 5.2 的得分与 Opus 5.5 的最佳得分（45.3% vs. 45.5%）大致相同，每项任务的成本约为后者的一半，且领先于 GPT-6 Astra 的最佳得分（42.8%），成本大致相当。在极高努力程度下，Fable 5.2 达到 49.1%，这是我们测试过的所有已发布模型中最高的得分。
WANDR (relaxed score)Accuracy vs. cost
WANDR（宽松评分） 准确率与成本
- Fable 5.2
- Opus 5.5
- Fable 5.1
304050607080900Score (%)125102050Cost per attempt (USD, log scale)LowMedHighXhighMax
30 40 50 60 70 80 90 0 分数 (%) 1 2 5 10 20 50 每次尝试成本 (美元，对数刻度) 低 中 高 极高 最大
Perplexity’s WANDR benchmark measures agents on large data-collection tasks. At Medium effort, Fable 5.2 matches Opus 5.5’s top score for about 40% of the cost per task.
Perplexity 的 WANDR 基准测试在大型数据收集任务上评估智能体。在中等努力程度下，Fable 5.2 的得分与 Opus 5.5 的最高得分持平，但每项任务的成本约为后者的 40%。
BenchCAD 2.1 (with tools)Accuracy vs. cost
BenchCAD 2.1（带工具） 准确率与成本
- Fable 5.2
- Opus 5.5
- Fable 5.1
020406080Score (0–100)0.030.100.30131030Cost per attempt (USD, log scale)LowMedHighXhighMax
0 20 40 60 80 分数 (0–100) 0.03 0.10 0.30 1 3 10 30 每次尝试成本 (美元，对数刻度) 低 中 高 极高 最大
BenchCAD 2.1 measures hardware engineering, specifically how closely a model’s designs match reference geometry and circuits, and is scored 0–100 with partial credit. At Max effort, Fable 5.2 scores 83.5, ahead of Opus 5.5 (65.3) and Fable 5.1 (44.4).
BenchCAD 2.1 衡量硬件工程能力，具体评估模型的设计与参考几何形状和电路的匹配程度，评分范围为 0–100，并包含部分得分。在最大努力程度下，Fable 5.2 得分为 83.5，领先于 Opus 5.5（65.3）和 Fable 5.1（44.4）。
Humanity’s Last Exam (with tools)Accuracy vs. cost
Humanity’s Last Exam（带工具） 准确率与成本
- Fable 5.2
- Opus 5.5
- Fable 5.1
506070800Score (%)0.100.200.50125Cost per attempt (USD, log scale)GPT-6 Astra (Max): 57.2%LowMedHighXhighMax
506070800得分（%）0.100.200.50125每次尝试成本（美元，对数刻度）GPT-6 Astra（最大值）：57.2%低中高高
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力