{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-07-28T05:20:57.982Z","headline":"OpenRouter 推出 Classifiers 测试版：自动标记 AI 请求的用途与成本归属","description":"OpenRouter 上线 Classifiers 测试版，允许用户通过自定义分类法（最多 8 个维度）自动标记每次 AI 请求的任务类型、部门归属、合规类别等信息。分类异步运行，不增加推理延迟；支持采样率控制成本，推荐使用 Gemini 3.5 Flash Lite 作为分类模型。标记结果写入日志，并可在 Activity Explorer 中按维度聚合分析模型使用分布与成本流向。","url":"https://www.aioga.com/news/cmryxzcmz06cbrolgm4zhtc1m/","mainEntityOfPage":"https://www.aioga.com/news/cmryxzcmz06cbrolgm4zhtc1m/","datePublished":"2026-07-24T00:00:00.000Z","dateModified":"2026-07-24T00:00:00.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://openrouter.ai/blog/announcements/classifiers","https://aihot.virxact.com/items/cmryxzcmz06cbrolgm4zhtc1m"],"canonicalUrl":"https://www.aioga.com/news/cmryxzcmz06cbrolgm4zhtc1m/","directAnswer":{"@type":"Answer","text":"OpenRouter 推出 Classifiers 测试版，用户可按自定义分类法为 AI 请求添加结构化标签，分类法最多包含八个维度。分类在请求完成后异步执行，结果写入日志，并可用于分析任务类型、模型使用和成本归属。","url":"https://www.aioga.com/news/cmryxzcmz06cbrolgm4zhtc1m/","dateCreated":"2026-07-24T00:00:00.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"openrouter.ai source article","url":"https://openrouter.ai/blog/announcements/classifiers","datePublished":"2026-07-24T00:00:00.000Z","provider":{"@type":"Organization","name":"openrouter.ai","url":"https://openrouter.ai/blog/announcements/classifiers"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmryxzcmz06cbrolgm4zhtc1m","datePublished":"2026-07-24T00:00:00.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmryxzcmz06cbrolgm4zhtc1m"}}],"aggregationSource":"OpenRouter：Announcements（RSS）","originalPublisher":{"name":"openrouter.ai","url":"https://openrouter.ai/blog/announcements/classifiers"},"article":{"id":"cmryxzcmz06cbrolgm4zhtc1m","slug":"cmryxzcmz06cbrolgm4zhtc1m","url":"https://www.aioga.com/news/cmryxzcmz06cbrolgm4zhtc1m/","title":"OpenRouter 推出 Classifiers 测试版：自动标记 AI 请求的用途与成本归属","title_en":"Classifiers： Track What Your Agents Do and What It Costs","summary":"OpenRouter 上线 Classifiers 测试版，允许用户通过自定义分类法（最多 8 个维度）自动标记每次 AI 请求的任务类型、部门归属、合规类别等信息。分类异步运行，不增加推理延迟；支持采样率控制成本，推荐使用 Gemini 3.5 Flash Lite 作为分类模型。标记结果写入日志，并可在 Activity Explorer 中按维度聚合分析模型使用分布与成本流向。","source":"OpenRouter：Announcements（RSS）","sourceUrl":"https://openrouter.ai/blog/announcements/classifiers","aiHotUrl":"https://aihot.virxact.com/items/cmryxzcmz06cbrolgm4zhtc1m","publishedAt":"2026-07-24T00:00:00.000Z","category":"产品更新","score":62,"selected":true,"articleBody":["You can now automatically classify your OpenRouter generations with structured metadata for AI usage reporting.","Every request carries information: the type of work, the level of complexity, which department it came from, whether it contains internal data it shouldn’t. Classifiers, now available in beta, give you that visibility. Define your criteria (task type, agent complexity, compliance category, cost center). A model of your choice tags each generation, or a sampled subset, against your taxonomy and write the results to your logs. You get continuous visibility into what your agents and users are doing, which models they’re using for different tasks, and where the costs go.","Create a classifier：https://openrouter.ai/workspaces/default/classifiers in your workspace settings, or read the docs：https://openrouter.ai/docs/guides/features/classifiers first.","A classifier is a small config with four parts: a taxonomy (up to eight dimensions, each with the values you choose), a classification prompt (instructions sent to the classifier model as a system message), a model to read each prompt and apply it, and a sampling rate . Classification runs asynchronously after each request completes, so it never adds latency to your inference path.","Choose from six preset templates, customize a template, or build your own from scratch.","Select your classification model. We recommend Gemini 3.5 Flash Lite：https://openrouter.ai/google/gemini-3.5-flash-lite for the best value: cheap, strong accuracy on structured output, good enough for most taxonomies. You can change the model at any time.","At high throughput, the cost of classifying every request adds up. Use the sampling rate to keep costs down. Run a high-fidelity compliance classifier at 100% while a broader cost-attribution classifier samples 10% of the same traffic, keeping costs proportional to the oversight you need.","Classifier outputs are coerced into structured formats, constrained to the dimensions and values you define. Every classified generation is tagged in your logs：https://openrouter.ai/logs, so you can filter requests by classification. For example, you can pull every request tagged department: legal or agent_complexity_difficulty_tier: complex_multistep . Each tagged generation’s detail panel breaks down classified dimensions and values.","You can also run a classifier on demand against any past generation to sanity-check a new taxonomy. Open it in your logs, pick a classifier, and see how it gets tagged.","Individual tags on generations answer “what was this request?” The Activity Explorer：https://openrouter.ai/activity/explore answers the aggregate questions: group your traffic by any classifier dimension to see which models are being used for each task type or level of agent complexity, and which departments or tasks drive the most spend.","Results are aggregated over time; watch patterns shift in your data and show stakeholders how your AI usage is governed. Classifier filters carry across the Activity tabs so you can see trends：https://openrouter.ai/activity/trends and guardrail：https://openrouter.ai/activity/guardrails enforcement by any classifier value.","Classifiers are available now in beta. Create a classifier：https://openrouter.ai/workspaces/default/classifiers in your workspace or read the docs：https://openrouter.ai/docs/guides/features/classifiers to learn more about taxonomy design, billing, and how classification works under the hood. Classifiers work even with input & output logging disabled.","Tell us what you think in #feedback：https://discord.gg/fVyRaUDgxW on Discord."],"articleImages":[{"sourceUrl":"https://openrouter.ai/blog/images/classifiers.png","alt":"Classifiers: Track What Your Agents Do and What It Costs","afterParagraph":0,"url":"/media/articles/cmryxzcmz06cbrolgm4zhtc1m/93797bfa1c641c45.png"},{"sourceUrl":"https://openrouter.ai/blog/images/classifiers-logs.png","alt":"Logs filtered by a classifier value, with a generation's detail panel showing its classified dimensions: difficulty tier, task family, and more","afterParagraph":8,"url":"/media/articles/cmryxzcmz06cbrolgm4zhtc1m/6a2ff161bebeff23.png"}],"mediaStatus":"ok","articleBodyZh":["您现在可以使用结构化元数据自动对您的 OpenRouter 生成内容进行分类，以用于 AI 使用报告。","每个请求都包含信息：工作类型、复杂程度、来源部门，以及是否包含不应使用的内部数据。现在可用的分类器（处于测试阶段）可以让您看到这些信息。定义您的标准（任务类型、代理复杂性、合规类别、成本中心）。您选择的模型会根据您设定的分类体系为每个生成内容或抽样子集打上标签，并将结果写入日志。您可以持续查看代理和用户的操作内容、他们在不同任务中使用的模型，以及费用的去向。","在工作区设置中创建分类器：https://openrouter.ai/workspaces/default/classifiers，或先阅读文档：https://openrouter.ai/docs/guides/features/classifiers。","分类器是一个包含四部分的小型配置：一个分类体系（最多八个维度，每个维度有您选择的值）、一个分类提示（作为系统消息发送给分类器模型的指令）、一个用于读取每个提示并应用分类的模型，以及一个抽样率。分类在每个请求完成后异步运行，因此不会增加推理路径的延迟。","从六个预设模板中选择、定制模板，或从零开始构建您自己的模板。","选择您的分类模型。我们推荐使用 Gemini 3.5 Flash Lite：https://openrouter.ai/google/gemini-3.5-flash-lite，以获得最佳性价比：成本低、在结构化输出上准确度高，对于大多数分类体系足够使用。您可以随时更改模型。","在高吞吐量情况下，对每个请求进行分类的成本会累积。使用抽样率可以控制成本。在进行高精度合规分类时，使用 100%，而更广泛的成本归因分类则对同一流量抽样 10%，保持成本与所需监督成正比。","分类器输出被强制转换为结构化格式，并受你定义的维度和值的约束。每次分类生成都会在你的日志中被标记：https://openrouter.ai/logs，因此你可以按分类过滤请求。例如，你可以提取所有标记为 department: legal 或 agent_complexity_difficulty_tier: complex_multistep 的请求。每个标记生成的详细面板都会拆解已分类的维度和值。","你还可以针对任何过去的生成按需运行分类器，以对新的分类法进行合理性检查。在日志中打开它，选择一个分类器，然后查看它是如何被标记的。","生成上的单个标签回答“这个请求是什么？”活动浏览器：https://openrouter.ai/activity/explore 回答汇总问题：按任意分类器维度分组你的流量，以查看每种任务类型或代理复杂度级别使用了哪些模型，以及哪些部门或任务产生了最多花费。","结果会随时间汇总；观察数据模式的变化，并向利益相关者展示你的 AI 使用情况如何被管理。分类器过滤器会在活动标签之间保持一致，因此你可以看到趋势：https://openrouter.ai/activity/trends 并通过任何分类器值执行护栏：https://openrouter.ai/activity/guardrails。","分类器现在以测试版形式提供。在你的工作区创建分类器：https://openrouter.ai/workspaces/default/classifiers 或阅读文档：https://openrouter.ai/docs/guides/features/classifiers 以了解更多关于分类法设计、计费以及分类器底层工作原理的信息。即使启用输入和输出日志记录，分类器也能工作。","在 Discord 的 #feedback 频道告诉我们你的想法：https://discord.gg/fVyRaUDgxW。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"OpenRouter 推出 Classifiers 测试版，用户可按自定义分类法为 AI 请求添加结构化标签，分类法最多包含八个维度。分类在请求完成后异步执行，结果写入日志，并可用于分析任务类型、模型使用和成本归属。","background":"该功能由分类法、分类提示词、分类模型和采样率四部分组成。用户可选择六种预设模板、修改模板或自行创建，并指定模型读取请求提示词，再按预设维度和值输出分类结果。","viewpoint":"Aioga 判断，Classifiers 的重点不是改变模型推理能力，而是为 AI 使用报告增加统一的元数据层。异步处理避免增加推理路径延迟，但分类质量仍可能受到分类法设计、提示词和所选模型影响。","implications":"企业可在日志中按部门、任务复杂度、合规类别或成本中心筛选请求，并在 Activity Explorer 中聚合观察模型使用分布与成本流向。值得关注的是，对全部请求分类会产生额外成本，采样率将影响覆盖程度。","nextStep":"Aioga 判断，测试阶段可先选取边界清晰的少量维度验证标签一致性，再根据监督需求设置采样率。合规分类可采用较高覆盖率，成本归属等宽泛分析则可尝试抽样，并持续检查日志中的分类结果。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-07-26T22:09:31.623Z","sourceHash":"34c5103d21a7f8ff","review":{"approved":true,"groundedness":94,"clarity":92,"duplicationRisk":12,"blockingIssues":[],"notes":["“分类质量仍可能受到分类法设计、提示词和所选模型影响”属于基于功能机制的合理判断，已明确标注为判断。","“先选取边界清晰的少量维度验证标签一致性”等内容属于编辑建议，已明确标注为 Aioga 判断，不构成事实性断言。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","low-source-overlap","no-html","independent-ai-review"]}},"tags":["产品更新","OpenRouter：Announcements（RSS）"],"translations":{"zh-CN":{"title":"OpenRouter 推出 Classifiers 测试版：自动标记 AI 请求的用途与成本归属","summary":"OpenRouter 上线 Classifiers 测试版，允许用户通过自定义分类法（最多 8 个维度）自动标记每次 AI 请求的任务类型、部门归属、合规类别等信息。分类异步运行，不增加推理延迟；支持采样率控制成本，推荐使用 Gemini 3.5 Flash Lite 作为分类模型。标记结果写入日志，并可在 Activity Explorer 中按维度聚合分析模型使用分布与成本流向。","category":"产品更新","source":"openrouter.ai","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter 推出 Classifiers 测试版：自动标记 AI 请求的用途与成本归属 - Aioga AI资讯","description":"OpenRouter 上线 Classifiers 测试版，允许用户通过自定义分类法（最多 8 个维度）自动标记每次 AI 请求的任务类型、部门归属、合规类别等信息。分类异步运行，不增加推理延迟；支持采样率控制成本，推荐使用 Gemini 3.5 Flash Lite 作为分类模型。标记结果写入日志，并可在 Activity Explorer 中按维度聚合分...","url":"https://www.aioga.com/news/cmryxzcmz06cbrolgm4zhtc1m/"},"en":{"title":"OpenRouter launches Classifiers beta: Automatically label the purpose and cost attribution of AI requests","summary":"OpenRouter has launched a beta version of Classifiers, allowing users to automatically label each AI request with information such as task type, department affiliation, and compliance category through a custom taxonomy (up to 8 dimensions). The classification runs asynchronously and does not add inference latency; it supports sampling rates to control costs and recommends using Gemini 3.5 Flash Lite as the classification model. The labeling results are written to the logs and can be aggregated by dimensions in the Activity Explorer to analyze model usage distribution and cost flow.","category":"Products","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter launches Classifiers beta: Automatically label the purpose and cost attribution of AI requests - Aioga AI News","description":"OpenRouter has launched a beta version of Classifiers, allowing users to automatically label each AI request with information such as task type, department affiliation, and complia...","url":"https://www.aioga.com/en/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:04:36.179Z"},"ja":{"title":"OpenRouter、Classifiersベータ版をリリース：AIリクエストの用途とコスト帰属を自動でラベル付け","summary":"OpenRouter は Classifiers テスト版を公開し、ユーザーがカスタム分類法（最大 8 つの次元）を通じて、各 AI リクエストのタスクタイプ、所属部門、コンプライアンスカテゴリなどの情報を自動的にタグ付けできるようになりました。分類は非同期で実行され、推論の遅延は増加しません。コスト制御のためにサンプリング率の設定をサポートしており、分類モデルには Gemini 3.5 Flash Lite の使用を推奨します。タグ付け結果はログに記録され、Activity Explorer で次元ごとに集計してモデル使用の分布とコストフローを分析できます。","category":"製品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter、Classifiersベータ版をリリース：AIリクエストの用途とコスト帰属を自動でラベル付け - Aioga AIニュース","description":"OpenRouter は Classifiers テスト版を公開し、ユーザーがカスタム分類法（最大 8 つの次元）を通じて、各 AI リクエストのタスクタイプ、所属部門、コンプライアンスカテゴリなどの情報を自動的にタグ付けできるようになりました。分類は非同期で実行され、推論の遅延は増加しません。コスト制御のためにサンプリング率の設定をサポートしており、分類モ...","url":"https://www.aioga.com/ja/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:04:45.772Z"},"ko":{"title":"OpenRouter, Classifiers 베타 버전 출시: AI 요청의 용도와 비용 귀속 자동 표시","summary":"OpenRouter에서 Classifiers 테스트 버전을 출시하여 사용자가 맞춤 분류법(최대 8개 차원)을 통해 각 AI 요청의 작업 유형, 부서 소속, 준수 카테고리 등의 정보를 자동으로 표시하도록 허용합니다. 분류는 비동기적으로 실행되어 추론 지연을 증가시키지 않습니다. 비용을 제어하기 위한 샘플링율을 지원하며, 분류 모델로 Gemini 3.5 Flash Lite 사용을 권장합니다. 마킹 결과는 로그에 기록되며, Activity Explorer에서 차원별로 모델 사용 분포와 비용 흐름을 집계하여 분석할 수 있습니다.","category":"제품 업데이트","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter, Classifiers 베타 버전 출시: AI 요청의 용도와 비용 귀속 자동 표시 - Aioga AI 뉴스","description":"OpenRouter에서 Classifiers 테스트 버전을 출시하여 사용자가 맞춤 분류법(최대 8개 차원)을 통해 각 AI 요청의 작업 유형, 부서 소속, 준수 카테고리 등의 정보를 자동으로 표시하도록 허용합니다. 분류는 비동기적으로 실행되어 추론 지연을 증가시키지 않습니다. 비용을 제어하기 위한 샘플링율을 지원하며, 분...","url":"https://www.aioga.com/ko/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:05:25.700Z"},"es":{"title":"OpenRouter lanza la versión de prueba de Classifiers: etiquetado automático del propósito y la asignación de costos de las solicitudes de IA","summary":"OpenRouter ha lanzado la versión de prueba de Classifiers, que permite a los usuarios etiquetar automáticamente el tipo de tarea, la pertenencia al departamento, la categoría de cumplimiento y otra información de cada solicitud de IA mediante una taxonomía personalizada (hasta 8 dimensiones). La clasificación se ejecuta de forma asíncrona, sin aumentar la latencia de inferencia; se admite el control de la tasa de muestreo para gestionar los costos, y se recomienda usar Gemini 3.5 Flash Lite como modelo de clasificación. Los resultados de las etiquetas se escriben en el registro y pueden agregarse y analizarse por dimensiones en Activity Explorer para examinar la distribución de uso del modelo y el flujo de costos.","category":"Productos","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter lanza la versión de prueba de Classifiers: etiquetado automático del propósito y la asignación de costos de las solicitudes de IA - Aioga Noticias de IA","description":"OpenRouter ha lanzado la versión de prueba de Classifiers, que permite a los usuarios etiquetar automáticamente el tipo de tarea, la pertenencia al departamento, la categoría de cu...","url":"https://www.aioga.com/es/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:05:27.544Z"},"fr":{"title":"OpenRouter lance la version test des Classifiers : marquage automatique de l'utilisation et des coûts des requêtes IA","summary":"OpenRouter lance la version test des Classifiers, permettant aux utilisateurs de marquer automatiquement chaque requête AI avec le type de tâche, l'appartenance au département, la catégorie de conformité, etc., via une taxonomie personnalisée (jusqu'à 8 dimensions). La classification fonctionne de manière asynchrone, sans ajouter de latence de raisonnement ; elle prend en charge le contrôle du taux d'échantillonnage pour gérer les coûts, et il est recommandé d'utiliser Gemini 3.5 Flash Lite comme modèle de classification. Les résultats des marquages sont enregistrés dans les journaux et peuvent être agrégés dans Activity Explorer pour analyser la distribution de l'utilisation du modèle et les flux de coûts selon chaque dimension.","category":"Produits","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter lance la version test des Classifiers : marquage automatique de l'utilisation et des coûts des requêtes IA - Aioga Actualités IA","description":"OpenRouter lance la version test des Classifiers, permettant aux utilisateurs de marquer automatiquement chaque requête AI avec le type de tâche, l'appartenance au département, la...","url":"https://www.aioga.com/fr/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:06:03.883Z"},"de":{"title":"OpenRouter startet die Testversion von Classifiers: Automatisches Kennzeichnen des Zwecks und der Kostenverteilung von KI-Anfragen","summary":"Die Testversion der Classifiers auf OpenRouter ist online und ermöglicht es den Nutzern, jede AI-Anfrage automatisch anhand einer benutzerdefinierten Taxonomie (bis zu 8 Dimensionen) zu kennzeichnen, z. B. nach Aufgabentyp, Abteilungszugehörigkeit und Compliance-Kategorie. Die Klassifizierung läuft asynchron und erhöht nicht die Inferenzverzögerung; die Stichprobenrate kann zur Kostenkontrolle angepasst werden. Es wird empfohlen, Gemini 3.5 Flash Lite als Klassifizierungsmodell zu verwenden. Die Markierungsergebnisse werden im Logbuch gespeichert und können im Activity Explorer nach Dimensionen aggregiert analysiert werden, um die Nutzung des Modells und die Kostenflüsse zu überwachen.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter startet die Testversion von Classifiers: Automatisches Kennzeichnen des Zwecks und der Kostenverteilung von KI-Anfragen - Aioga KI-News","description":"Die Testversion der Classifiers auf OpenRouter ist online und ermöglicht es den Nutzern, jede AI-Anfrage automatisch anhand einer benutzerdefinierten Taxonomie (bis zu 8 Dimensione...","url":"https://www.aioga.com/de/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:06:05.412Z"},"pt-BR":{"title":"OpenRouter lança versão de teste dos Classifiers: marca automaticamente o uso e a alocação de custos dos pedidos de IA","summary":"OpenRouter lançou a versão de teste dos Classifiers, permitindo que os usuários etiquetem automaticamente cada solicitação de IA por tipo de tarefa, departamento, categoria de conformidade e outras informações através de uma taxonomia personalizada (até 8 dimensões). A classificação é executada de forma assíncrona, sem aumentar a latência da inferência; suporta controle de taxa de amostragem para gerenciar custos, sendo recomendado o uso do Gemini 3.5 Flash Lite como modelo de classificação. Os resultados da marcação são gravados nos logs e podem ser agregados e analisados por dimensão no Activity Explorer para examinar a distribuição de uso do modelo e o fluxo de custos.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter lança versão de teste dos Classifiers: marca automaticamente o uso e a alocação de custos dos pedidos de IA - Aioga Notícias de IA","description":"OpenRouter lançou a versão de teste dos Classifiers, permitindo que os usuários etiquetem automaticamente cada solicitação de IA por tipo de tarefa, departamento, categoria de conf...","url":"https://www.aioga.com/pt-BR/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:06:39.106Z"},"ru":{"title":"OpenRouter выпустил бета-версию Classifiers: автоматическая маркировка назначения и распределения расходов запросов ИИ","summary":"На OpenRouter запущена бета-версия Classifiers, которая позволяет пользователям автоматически помечать тип задачи, принадлежность к отделу, категорию соответствия и другую информацию для каждого AI-запроса с помощью пользовательской таксономии (до 8 измерений). Классификация выполняется асинхронно и не увеличивает задержку вывода; поддерживается контроль частоты выборки для управления стоимостью. Рекомендуется использовать модель Gemini 3.5 Flash Lite в качестве классификатора. Результаты пометок записываются в журнал и могут быть агрегированы по измерениям в Activity Explorer для анализа распределения использования модели и потоков затрат.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter выпустил бета-версию Classifiers: автоматическая маркировка назначения и распределения расходов запросов ИИ - Aioga Новости ИИ","description":"На OpenRouter запущена бета-версия Classifiers, которая позволяет пользователям автоматически помечать тип задачи, принадлежность к отделу, категорию соответствия и другую информац...","url":"https://www.aioga.com/ru/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:06:43.705Z"},"ar":{"title":"أطلقت OpenRouter النسخة التجريبية من المصنفات: وضع علامات تلقائية على استخدامات وتكاليف طلبات الذكاء الاصطناعي","summary":"أصدر OpenRouter النسخة التجريبية من المصنفات التي تسمح للمستخدمين بتصنيف نوع مهمة كل طلب AI، والانتماء للقسم، والفئة الامتثالية، وغيرها من المعلومات تلقائيًا من خلال تصنيفات مخصصة (بحد أقصى 8 أبعاد). يتم تشغيل التصنيف بشكل غير متزامن، دون زيادة تأخير الاستدلال؛ ويدعم التحكم في معدل العينات لتقليل التكلفة، ويوصى باستخدام Gemini 3.5 Flash Lite كنموذج تصنيف. تُكتب نتائج التصنيف في السجل، ويمكن تجميعها وتحليلها حسب البُعد في مستكشف النشاط لفهم توزيع استخدام النموذج وتدفق التكاليف.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"أطلقت OpenRouter النسخة التجريبية من المصنفات: وضع علامات تلقائية على استخدامات وتكاليف طلبات الذكاء الاصطناعي - Aioga أخبار الذكاء الاصطناعي","description":"أصدر OpenRouter النسخة التجريبية من المصنفات التي تسمح للمستخدمين بتصنيف نوع مهمة كل طلب AI، والانتماء للقسم، والفئة الامتثالية، وغيرها من المعلومات تلقائيًا من خلال تصنيفات مخصصة...","url":"https://www.aioga.com/ar/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:07:24.088Z"},"hi":{"title":"OpenRouter ने Classifiers टेस्ट संस्करण लॉन्च किया: AI अनुरोधों के उपयोग और लागत के जिम्मेदार को स्वचालित रूप से चिह्नित करता है","summary":"OpenRouter पर Classifiers बीटा संस्करण लॉन्च हुआ है, जो उपयोगकर्ताओं को कस्टम टैक्सोनॉमी (अधिकतम 8 आयाम) के माध्यम से प्रत्येक AI अनुरोध के कार्य प्रकार, विभागीय स्वामित्व, अनुपालन श्रेणी आदि की जानकारी स्वचालित रूप से चिह्नित करने की अनुमति देता है। वर्गीकरण असिंक्रोनस तरीके से चलता है, जिससे अनुमान में विलंब नहीं बढ़ता; लागत को नियंत्रित करने के लिए सैंपल दर का समर्थन करता है, और वर्गीकरण मॉडल के रूप में Gemini 3.5 Flash Lite का उपयोग करने की सिफारिश की जाती है। चिह्नित परिणाम लॉग में लिखे जाते हैं, और Activity Explorer में आयामों के अनुसार मॉडल उपयोग वितरण और लागत प्रवाह का समेकित विश्लेषण किया जा सकता है।","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter ने Classifiers टेस्ट संस्करण लॉन्च किया: AI अनुरोधों के उपयोग और लागत के जिम्मेदार को स्वचालित रूप से चिह्नित करता है - Aioga AI समाचार","description":"OpenRouter पर Classifiers बीटा संस्करण लॉन्च हुआ है, जो उपयोगकर्ताओं को कस्टम टैक्सोनॉमी (अधिकतम 8 आयाम) के माध्यम से प्रत्येक AI अनुरोध के कार्य प्रकार, विभागीय स्वामित्व, अनुपालन...","url":"https://www.aioga.com/hi/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:07:26.299Z"},"it":{"title":"OpenRouter lancia la versione beta dei Classifiers: etichettatura automatica dell'uso e degli oneri dei richieste AI","summary":"OpenRouter ha lanciato la versione di test dei Classifiers, che consente agli utenti di etichettare automaticamente ogni richiesta AI con il tipo di attività, il dipartimento di appartenenza, la categoria di conformità e altre informazioni tramite tassonomie personalizzate (fino a 8 dimensioni). La classificazione viene eseguita in modo asincrono, senza aumentare la latenza dell'inferenza; supporta la regolazione del tasso di campionamento per controllare i costi, e si consiglia di utilizzare Gemini 3.5 Flash Lite come modello di classificazione. I risultati delle etichette vengono scritti nei log e possono essere aggregati per dimensione in Activity Explorer per analizzare la distribuzione dell'uso del modello e il flusso dei costi.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter lancia la versione beta dei Classifiers: etichettatura automatica dell'uso e degli oneri dei richieste AI - Aioga Notizie IA","description":"OpenRouter ha lanciato la versione di test dei Classifiers, che consente agli utenti di etichettare automaticamente ogni richiesta AI con il tipo di attività, il dipartimento di ap...","url":"https://www.aioga.com/it/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:08:06.682Z"},"nl":{"title":"OpenRouter lanceert bèta van Classifiers: automatisch het doel en de kostenverdeling van AI-verzoeken labelen","summary":"OpenRouter lanceert de bètaversie van Classifiers, waarmee gebruikers elk AI-verzoek automatisch kunnen labelen op taaktype, afdelingslidmaatschap, compliancecategorieën en andere informatie via een zelf aangepaste taxonomie (maximaal 8 dimensies). Classificatie wordt asynchroon uitgevoerd, waardoor de vertragingsrisico's bij inferentie niet toenemen; ondersteuning voor sampling-rate maakt kostbeheersing mogelijk. Het wordt aanbevolen Gemini 3.5 Flash Lite als classificatiemodel te gebruiken. De labelresultaten worden in de log geschreven en kunnen in Activity Explorer per dimensie worden geaggregeerd om het gebruik van het model en de kostenstroom te analyseren.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter lanceert bèta van Classifiers: automatisch het doel en de kostenverdeling van AI-verzoeken labelen - Aioga AI-nieuws","description":"OpenRouter lanceert de bètaversie van Classifiers, waarmee gebruikers elk AI-verzoek automatisch kunnen labelen op taaktype, afdelingslidmaatschap, compliancecategorieën en andere...","url":"https://www.aioga.com/nl/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:08:03.086Z"},"tr":{"title":"OpenRouter, Classifiers Beta Sürümünü Tanıttı: AI Taleplerinin Kullanım Amacı ve Maliyet Sahipliğini Otomatik Olarak İşaretleme","summary":"OpenRouter, Kullanıcılara her AI isteğinin görev türünü, birim aidiyetini, uyumluluk kategorisini vb. bilgilerini en fazla 8 boyutta özelleştirilmiş taksonomi aracılığıyla otomatik olarak işaretleme imkanı veren Classifiers test sürümünü yayına aldı. Sınıflandırma asenkron olarak çalışır ve çıkarım gecikmesini artırmaz; maliyeti kontrol etmek için örnekleme oranı desteği bulunur ve sınıflandırma modeli olarak Gemini 3.5 Flash Lite kullanılması önerilir. İşaretleme sonuçları günlük kaydına yazılır ve Activity Explorer'da boyutlara göre toplanıp analiz edilerek model kullanım dağılımları ve maliyet akışları incelenebilir.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter, Classifiers Beta Sürümünü Tanıttı: AI Taleplerinin Kullanım Amacı ve Maliyet Sahipliğini Otomatik Olarak İşaretleme - Aioga AI Haberleri","description":"OpenRouter, Kullanıcılara her AI isteğinin görev türünü, birim aidiyetini, uyumluluk kategorisini vb. bilgilerini en fazla 8 boyutta özelleştirilmiş taksonomi aracılığıyla otomatik...","url":"https://www.aioga.com/tr/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:08:47.327Z"},"vi":{"title":"OpenRouter ra mắt phiên bản thử nghiệm Classifiers: Tự động gán nhãn mục đích và chi phí của các yêu cầu AI","summary":"OpenRouter ra mắt phiên bản thử nghiệm Classifiers, cho phép người dùng tự động đánh dấu loại tác vụ, bộ phận thuộc về, và loại tuân thủ của mỗi yêu cầu AI thông qua phân loại tùy chỉnh (tối đa 8 chiều). Việc phân loại chạy bất đồng bộ, không làm tăng độ trễ suy luận; hỗ trợ kiểm soát tỷ lệ mẫu để giảm chi phí, khuyến nghị sử dụng Gemini 3.5 Flash Lite làm mô hình phân loại. Kết quả đánh dấu được ghi vào nhật ký và có thể được tổng hợp phân tích phân phối sử dụng mô hình và dòng chi phí theo các chiều trong Activity Explorer.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter ra mắt phiên bản thử nghiệm Classifiers: Tự động gán nhãn mục đích và chi phí của các yêu cầu AI - Tin tức AI Aioga","description":"OpenRouter ra mắt phiên bản thử nghiệm Classifiers, cho phép người dùng tự động đánh dấu loại tác vụ, bộ phận thuộc về, và loại tuân thủ của mỗi yêu cầu AI thông qua phân loại tùy...","url":"https://www.aioga.com/vi/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:08:48.514Z"},"id":{"title":"OpenRouter meluncurkan versi uji coba Classifiers: Menandai secara otomatis tujuan dan alokasi biaya permintaan AI","summary":"OpenRouter meluncurkan versi percobaan Classifiers, memungkinkan pengguna untuk secara otomatis menandai jenis tugas, afiliasi departemen, kategori kepatuhan, dan informasi lainnya untuk setiap permintaan AI melalui taksonomi kustom (maksimal 8 dimensi). Klasifikasi berjalan secara asinkron, tanpa menambah latensi inferensi; mendukung kontrol tingkat sampling untuk mengelola biaya, disarankan menggunakan Gemini 3.5 Flash Lite sebagai model klasifikasi. Hasil penandaan ditulis ke log, dan dapat dianalisis secara agregat di Activity Explorer berdasarkan dimensi untuk distribusi penggunaan model dan aliran biaya.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter meluncurkan versi uji coba Classifiers: Menandai secara otomatis tujuan dan alokasi biaya permintaan AI - Berita AI Aioga","description":"OpenRouter meluncurkan versi percobaan Classifiers, memungkinkan pengguna untuk secara otomatis menandai jenis tugas, afiliasi departemen, kategori kepatuhan, dan informasi lainnya...","url":"https://www.aioga.com/id/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:09:22.149Z"},"th":{"title":"OpenRouter เปิดตัว Classifiers รุ่นทดสอบ: ทำเครื่องหมายการใช้งานและค่าใช้จ่ายของคำขอ AI อัตโนมัติ","summary":"OpenRouter เปิดตัวรุ่นทดสอบ Classifiers ซึ่งอนุญาตให้ผู้ใช้สามารถทำการแท็กประเภทงาน หน่วยงานที่เกี่ยวข้อง หมวดหมู่การปฏิบัติตามกฎระเบียบ และข้อมูลอื่น ๆ ของ AI ทุกคำขอโดยอัตโนมัติ ผ่านการจำแนกประเภทที่กำหนดเอง (สูงสุด 8 มิติ) การจำแนกทำงานแบบอะซิงโครนัส ไม่เพิ่มความหน่วงในการให้เหตุผล; รองรับการควบคุมอัตราการสุ่มเพื่อควบคุมค่าใช้จ่าย แนะนำให้ใช้ Gemini 3.5 Flash Lite เป็นโมเดลการจำแนก ผลลัพธ์การแท็กจะถูกเขียนลงในบันทึกและสามารถวิเคราะห์การใช้โมเดลและการไหลของค่าใช้จ่ายตามมิติผ่าน Activity Explorer","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter เปิดตัว Classifiers รุ่นทดสอบ: ทำเครื่องหมายการใช้งานและค่าใช้จ่ายของคำขอ AI อัตโนมัติ - ข่าว AI Aioga","description":"OpenRouter เปิดตัวรุ่นทดสอบ Classifiers ซึ่งอนุญาตให้ผู้ใช้สามารถทำการแท็กประเภทงาน หน่วยงานที่เกี่ยวข้อง หมวดหมู่การปฏิบัติตามกฎระเบียบ และข้อมูลอื่น ๆ ของ AI ทุกคำขอโดยอัตโนมัติ...","url":"https://www.aioga.com/th/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:09:34.554Z"},"pl":{"title":"OpenRouter wprowadza wersję testową Classifiers: automatyczne oznaczanie celu i kosztów żądań AI","summary":"OpenRouter uruchomił wersję testową klasyfikatorów, pozwalając użytkownikom automatycznie oznaczać typ zadania, przynależność do działu, kategorię zgodności i inne informacje dla każdego żądania AI za pomocą niestandardowej taksonomii (maksymalnie 8 wymiarów). Klasyfikacja działa asynchronicznie, nie wydłużając czasu inferencji; wspiera kontrolę kosztów poprzez regulację próbkowania, zaleca się użycie Gemini 3.5 Flash Lite jako modelu klasyfikacyjnego. Wyniki oznaczania są zapisywane w logach i mogą być analizowane w Activity Explorer, agregując dane według wymiarów w celu przeanalizowania dystrybucji użytkowania modelu i przepływu kosztów.","category":"产品更新","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter wprowadza wersję testową Classifiers: automatyczne oznaczanie celu i kosztów żądań AI - Aioga Wiadomości AI","description":"OpenRouter uruchomił wersję testową klasyfikatorów, pozwalając użytkownikom automatycznie oznaczać typ zadania, przynależność do działu, kategorię zgodności i inne informacje dla k...","url":"https://www.aioga.com/pl/news/cmryxzcmz06cbrolgm4zhtc1m/","contentTranslated":true,"sourceHash":"15bd8654fc1e8da2","translatedAt":"2026-07-26T03:10:15.372Z"}}}}