Anthropic 发布 Claude Fable 5.1 与 Claude Mythos 5.1：Terminal-Bench-Science 得分 52.6%，缓存读取降价 75%

Anthropic 在 2026 年 6 月发布 Fable 5 系列三个月后，发布了 Claude Fable 5.1 和 Claude Mythos 5.1。这两者是相同的基础模型，但有不同的安全保护层。Fable 5.1 可作为 claude-fable-5-1 普遍使用；Mythos 5.1 限制在经过审查的机构内部使用。两者都具备 100 万令牌的上下文窗口和最大 12.8 万输出令牌，始终启用自适应思考。主要能力指标在 Terminal-Bench-Science 0.1 上为 52.6%，相比 Fable 5 的 24.7% 和 Opus 5 的 29.0%。主要商业指标是缓存读取费用减少 75%，从每百万令牌 1.00 美元降至 0.25 美元，Anthropic 测量得出在典型工作负载中成本约降低 25%，在自主任务中最高可降至 45%。基础输入和输出定价保持不变，为每百万令牌 10 美元和 50 美元。

是的，Claude Fable 5.1：https://www.anthropic.com/claude-fable-and-mythos-5-1 在 Claude API、Amazon Bedrock、AWS 上的 Claude 平台、谷歌云和 Microsoft Foundry 上可作为 claude-fable-5-1 普遍使用。Claude Mythos 5.1 不可用，它仅限于 Project Glasswing 内经过审查的美国机构使用：https://anthropic.com/glasswing。

在 Terminal-Bench-Science 0.1 上，这是一个自主科研基准测试，Fable 5.1 得分为 52.6%，相比 Opus 5 的 29.0%，Fable 5 的 24.7% 和 GPT-5.6 Sol 的 22.4%。Anthropic 报告每个模型的标准误差为 3.5 至 4.5 分，因此应注意分数差距，而不仅仅是排名。

在 Terminal-Bench 4.0 上，Fable 5.1 达到 55.8%，Mythos 5.1 达到 60.9%。两个相同模型之间的差距是安全干预的成本，这是一个异常诚实的披露。其他方面：CursorBench 3.2.0 为 73.4%，Humanity’s Last Exam 在无工具下为 60.9%，有工具为 65.0%，AutomationBench 为 31.4%，OSWorld 2.0 严格为 41.7%，GDPval-AA v2 为 1853。

基础输入和输出定价保持不变。缓存读取减少 75%，从每百万令牌 1.00 美元降至 0.25 美元，相对于其他每个 Claude 模型 0.1 倍基准输入，仅为 0.025 倍。Anthropic 测量得出，在典型工作负载中成本约降低 25%，在上下文密集型自主任务中最高可降至约 45%。批量处理为每百万令牌 5 美元和 25 美元。

附加更改：每条消息的工作量、以轮次为范围的系统消息，以及 thinking.display: "updates" 都处于标头下的测试阶段。内容来源不可选择，所有输出都有统计文本水印，文件具有 C2PA 证书。

Anthropic 还记录了真实的回归。并行工具调用的变化更大，因此代理循环可能每轮只发出一次调用，而 Fable 5 则批处理了多个调用。模型讲述得更少，低工作量下更常从记忆中回答，并且偏好整个文件重写而非局部编辑。

网络安全防护现在允许发现漏洞，但不允许开发利用，减少了 Claude Code 每次会话约 60% 的干预。生物学安全防护在 benign 请求上触发频率降低 85%。渗透测试、漏洞生成和基于二进制的漏洞扫描仍会重定向到 Opus。

在研究方面，Mythos 5.1 在 12 个靶点上设计蛋白质结合剂的命中率约为 50%，而正常水平为 10% 到 15%；Fable 5.1 构建了金星高程图：https://zenodo.org/records/22164484，分辨率为 2 到 3 公里；自定义 GPU 内核将七个开源基因组学模型的速度提升了最多 2.5 倍。

需要与我们合作以推广您的 GitHub 仓库或 Hugging Face 页面或产品发布或网络研讨会等？请联系：https://forms.gle/wbash1wF6efRj8G58

Asif Razzaq 是 Marktechpost Media Inc. 的首席执行官。作为一名具有远见的企业家和工程师，Asif 致力于利用人工智能的潜力造福社会。他最近的努力是推出人工智能媒体平台 Marktechpost，该平台以对机器学习和深度学习新闻的深入报道而脱颖而出，这些报道既具有技术性又易于广泛读者理解。该平台每月浏览量超过 200 万，说明其在观众中非常受欢迎。

Anthropic has released Claude Fable 5.1 and Claude Mythos 5.1, three months after the Fable 5 line shipped in June 2026. The two are the same underlying model behind different safeguard layers. Fable 5.1 is generally available as claude-fable-5-1 ; Mythos 5.1 stays restricted to vetted organizations. Both carry a 1M token context window and 128K max output tokens, with adaptive thinking always on. The headline capability number is 52.6% on Terminal-Bench-Science 0.1, against 24.7% for Fable 5 and 29.0% for Opus 5. The headline commercial number is a 75% cut to cache reads, from $1.00 to $0.25 per million tokens, which Anthropic measures as roughly 25% lower cost on typical workloads and up to 45% on agentic ones. Base input and output pricing is unchanged at $10 and $50 per million.

Yes, Claude Fable 5.1：https://www.anthropic.com/claude-fable-and-mythos-5-1 is generally available as claude-fable-5-1 on the Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud, and Microsoft Foundry. Claude Mythos 5.1 is not: it is restricted to vetted US organizations inside Project Glasswing：https://anthropic.com/glasswing.

On Terminal-Bench-Science 0.1, an agentic scientific research benchmark, Fable 5.1 scores 52.6% against 29.0% for Opus 5, 24.7% for Fable 5, and 22.4% for GPT-5.6 Sol. Anthropic reports a standard error of 3.5 to 4.5 points per model, so treat the margin, not the ranking, with care.

On Terminal-Bench 4.0, Fable 5.1 reaches 55.8% and Mythos 5.1 reaches 60.9%. The gap between two identical models is the cost of safeguard interventions, which is an unusually honest disclosure. Elsewhere: CursorBench 3.2.0 at 73.4%, Humanity’s Last Exam at 60.9% without tools and 65.0% with tools, AutomationBench at 31.4%, OSWorld 2.0 at 41.7% strict, and GDPval-AA v2 at 1853.

Base input and output pricing is unchanged. Cache reads drop 75%, from $1.00 to $0.25 per million tokens, which is 0.025 times base input against 0.1 on every other Claude model. Anthropic measures roughly 25% lower cost on typical workloads and up to about 45% on context-heavy agentic ones. Batch processing is $5 and $25 per million tokens.

Additive changes: per-message effort, turn-scoped system messages, and thinking.display: "updates" are all in beta behind headers. Content provenance is not optional, with a statistical text watermark on all output and C2PA credentials on files.

Anthropic also documents real regressions. Parallel tool calling is more variable, so agent loops may issue one call per turn where Fable 5 batched several. The model narrates less, answers from memory more often at low effort, and prefers whole-file rewrites over targeted edits.

Cyber safeguards now permit vulnerability discovery but not exploit development, cutting interventions in Claude Code by roughly 60% per session. Biology safeguards fire 85% less often on benign requests. Penetration testing, exploit generation, and binary-based vulnerability scanning still redirect to Opus.

On research, Mythos 5.1 designed protein binders with roughly 50% hit rate across 12 targets against a 10 to 15% norm, Fable 5.1 built a Venus elevation map：https://zenodo.org/records/22164484 at 2 to 3 km resolution, and custom GPU kernels sped up seven open-source genomics models by up to 2.5x.

Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us ：https://forms.gle/wbash1wF6efRj8G58

Asif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.