[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-claude-fable-5-mythos-auto-routing":3,"topics-all":36,"news-related-aa29e468-6de8-417b-99f3-efe024635249":55},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":23,"news_slug":29,"published_at":30,"created_at":31,"modified_at":32,"is_published":33,"publish_type":34,"image_url":13,"view_count":35},"aa29e468-6de8-417b-99f3-efe024635249","Claude Fable 5 落地：Mythos 级能力首次走向大众，\"自动路由\"改写安全发布规则","6月9日，Anthropic 正式发布 Claude Fable 5，这是 Mythos 级模型首次向公众开放。与此前被严格限定的 Mythos 5 不同，Fable 5 走的是一条\"自动降级\"路线：模型本身不变，但在网络安全、生物学、化学、模型蒸馏等高风险主题上自动转回 Claude Opus 4.8 回答。Anthropic 表示这套保险机制平均不到 5% 的会话会触发，可以同时保住能力上限和合规底线。\n\n能力上，Fable 5 在软件工程、知识工作、视觉、长程自主任务上都达到 SOTA。Stripe 内部测试显示，它在 5000 万行 Ruby 代码上完成了一次代码库级迁移——按传统方式要一个团队干两个月。在 Cognition 的 FrontierCode 评测中，它在中等推理强度下仍是得分最高的 frontier 模型。在 Hebbia 金融基准上，文档推理、图表分析、问题求解三项都拿下第一名。\n\n定价方面，Fable 5 与 Mythos 5 同价：每百万输入 token 10 美元、输出 50 美元，是 Mythos Preview 的一半。订阅用户从 6 月 9 日到 6 月 22 日可在 Pro\u002FMax\u002FTeam\u002FEnterprise 计划里免费用，之后切到按量计费。\n\n更值得关注的其实是\"路由式安全\"这个范式：不再靠拒答、不再靠不发布，而是把强模型和低风险模型绑成一个路由层，让大多数用户根本感受不到差异。如果这套打法稳定，Anthropic 给整个行业立了一个\"强模型如何安全上线\"的样板。","https:\u002F\u002Fwww.anthropic.com\u002Fnews\u002Fclaude-fable-5-mythos-5","1fa87d30-d9f3-4752-b3be-0373933b3aaf",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"1fcfaaf2-67de-43d3-9e35-5784852fec60","ai-safety",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"23544f6a-eea1-4f05-aa8d-749ca862d5d2","anthropic",{"id":18,"name":19,"slug":19,"description":13,"color":13},"dca4d0ab-7994-43a7-839e-7756fc77344a","claude",{"id":21,"name":22,"slug":22,"description":13,"color":13},"7e89b5cc-57db-4f37-bc6d-28919a73931c","model-release",[24],{"id":25,"lang":26,"title":27,"summary":28,"content":13},"d27e9188-b8b9-4a82-9814-867c90950fae","en","Claude Fable 5 lands: Mythos-class power goes mainstream","On June 9 Anthropic officially released Claude Fable 5 — the first Mythos-class model to be opened to the public. Unlike the strictly gated Mythos 5, Fable 5 takes an \"automatic degradation\" path: the model itself is unchanged, but on high-risk topics (cybersecurity, biology, chemistry, model distillation, etc.) it automatically falls back to Claude Opus 4.8 to answer. Anthropic says this safety mechanism triggers on fewer than 5% of sessions on average, holding both the capability ceiling and the compliance floor.\n\nIn capability, Fable 5 hits SOTA across software engineering, knowledge work, vision, and long-horizon autonomous tasks. Stripe's internal test shows it completed a codebase-level migration across 50 million lines of Ruby code — work that would have taken a traditional team two months. On Cognition's FrontierCode benchmark, it remains the highest-scoring frontier model at moderate reasoning intensity. On Hebbia's financial benchmark, it takes first place in document reasoning, chart analysis, and problem solving.\n\nOn pricing, Fable 5 matches Mythos 5: $10 per million input tokens, $50 per million output, half the price of Mythos Preview. Subscribers can use it free on Pro\u002FMax\u002FTeam\u002FEnterprise plans from June 9 to June 22, then it switches to usage-based billing.\n\nThe more noteworthy thing is actually the \"routed safety\" paradigm: no longer relying on refusal, no longer relying on non-release, but tying a strong model and a low-risk model into a routing layer, so most users don't feel the difference. If this approach stabilizes, Anthropic has set a template for the entire industry on \"how to safely launch a strong model.\"","claude-fable-5-mythos-auto-routing","2026-06-09T03:00:00Z","2026-06-10T02:12:05.655687Z","2026-08-19T02:08:40.142862Z",true,"agent",169,[37,46],{"slug":38,"tag_slug":38,"title_zh":39,"title_en":40,"intro_zh":41,"intro_en":42,"id":43,"is_active":33,"created_at":44,"modified_at":45},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":47,"tag_slug":47,"title_zh":48,"title_en":49,"intro_zh":50,"intro_en":51,"id":52,"is_active":33,"created_at":53,"modified_at":54},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":56},[57,62,67,72,77,82],{"id":58,"title":59,"news_slug":60,"published_at":61},"63b28ef5-3ffa-4897-88d2-5dcd7fb678b5","Claude Fable 5.1 发布:缓存读取降价 75%,Agent 科研基准翻倍","claude-fable-5-1-mythos-release","2026-09-02T13:20:00+00:00",{"id":63,"title":64,"news_slug":65,"published_at":66},"f3d17d45-e1a8-4a1b-9449-6813aff06e49","Anthropic 让 Claude 自己修对齐:10 类失败全部见效,还超过人类研究员","claude-automated-alignment-researchers","2026-08-29T13:05:00+00:00",{"id":68,"title":69,"news_slug":70,"published_at":71},"97c97b9c-e6e4-4982-aa57-0c0da814fb19","Anthropic 的欧盟答卷四小时即被撕开：Claude 文本水印为什么怕改写","claude-synthid-70-percent-threshold-bypass","2026-08-21T08:00:00+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"470b8663-3916-4bc5-ac3c-c592487c2873","Claude水印官宣4小时被破:开源去除工具走红,水印军备竞赛开场","claude-watermark-removal-tool","2026-08-20T19:30:00+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"6b07def3-2b1c-4fe9-9922-0dd0038c149c","Anthropic 风险报告更新:Threat Model 1 升至「低」,Threat Model 2 维持「低」但信心下降","anthropic-risk-report-august-2026-update","2026-08-19T03:00:00+00:00",{"id":83,"title":84,"news_slug":85,"published_at":86},"a7f4cfad-874e-42b0-a84b-bd0ec57e8fdc","Anthropic 给 Claude 文本上不可见水印,接 SynthID-Text 走全球合规","anthropic-claude-invisible-text-watermark","2026-08-18T03:30:00+00:00"]