{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-09-03T00:00:45.782Z","headline":"GitHub Copilot 如何在不牺牲任务质量的前提下降低 AI 编码成本","description":"GitHub 工程师 Erik Kristensen 分享了 Copilot 降本的四项改动：选择性压缩工具输出、移除 view 工具行号前缀（线下推理成本降约 5%，线上用户日均推理成本降约 3%）、压缩 task-tool 提示词（每轮省约 1300 token，每活跃小时归一化成本降 2.9%）、后台任务完成后直接交付结果（AI Credits 用量降约 2.3%）。","url":"https://www.aioga.com/news/cmtkfkfvl03garobqyr0vwpth/","mainEntityOfPage":"https://www.aioga.com/news/cmtkfkfvl03garobqyr0vwpth/","datePublished":"2026-09-02T18:00:00.000Z","dateModified":"2026-09-02T18:00:00.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://github.blog/ai-and-ml/github-copilot/how-we-make-ai-coding-more-cost-efficient-without-sacrificing-task-quality","https://aihot.virxact.com/items/cmtkfkfvl03garobqyr0vwpth"],"canonicalUrl":"https://www.aioga.com/news/cmtkfkfvl03garobqyr0vwpth/","directAnswer":{"@type":"Answer","text":"GitHub Blog介绍了Copilot在不牺牲任务质量的前提下降低AI编码成本的四项改动：选择性压缩工具输出、移除view工具行号前缀、压缩task-tool提示词，以及后台任务完成后直接交付结果。","url":"https://www.aioga.com/news/cmtkfkfvl03garobqyr0vwpth/","dateCreated":"2026-09-02T18:00:00.000Z","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":"GitHub Blog source article","url":"https://github.blog/ai-and-ml/github-copilot/how-we-make-ai-coding-more-cost-efficient-without-sacrificing-task-quality","datePublished":"2026-09-02T18:00:00.000Z","provider":{"@type":"Organization","name":"GitHub Blog","url":"https://github.blog/ai-and-ml/github-copilot/how-we-make-ai-coding-more-cost-efficient-without-sacrificing-task-quality"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmtkfkfvl03garobqyr0vwpth","datePublished":"2026-09-02T18:00:00.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmtkfkfvl03garobqyr0vwpth"}}],"aggregationSource":"GitHub Blog","originalPublisher":{"name":"GitHub Blog","url":"https://github.blog/ai-and-ml/github-copilot/how-we-make-ai-coding-more-cost-efficient-without-sacrificing-task-quality"},"geoDeepAnswer":null,"article":{"id":"cmtkfkfvl03garobqyr0vwpth","slug":"cmtkfkfvl03garobqyr0vwpth","url":"https://www.aioga.com/news/cmtkfkfvl03garobqyr0vwpth/","title":"GitHub Copilot 如何在不牺牲任务质量的前提下降低 AI 编码成本","title_en":"","summary":"GitHub 工程师 Erik Kristensen 分享了 Copilot 降本的四项改动：选择性压缩工具输出、移除 view 工具行号前缀（线下推理成本降约 5%，线上用户日均推理成本降约 3%）、压缩 task-tool 提示词（每轮省约 1300 token，每活跃小时归一化成本降 2.9%）、后台任务完成后直接交付结果（AI Credits 用量降约 2.3%）。","source":"GitHub Blog","sourceUrl":"https://github.blog/ai-and-ml/github-copilot/how-we-make-ai-coding-more-cost-efficient-without-sacrificing-task-quality","aiHotUrl":"https://aihot.virxact.com/items/cmtkfkfvl03garobqyr0vwpth","publishedAt":"2026-09-02T18:00:00.000Z","category":"行业动态","score":72,"selected":true,"articleBody":["Bring agentic workflows to your terminal with GitHub Copilot CLI >：https://github.com/features/copilot/cli","@erik-krogh：https://github.com/erik-krogh","Erik Krogh Kristensen is a Staff Software Engineer at GitHub building at the intersection of AI, coding, and security. In the last few years at GitHub he's worked on an assortment of AI products such as Copilot Autofix, Copilot Code Review, and most recently Copilot CLI. He's a trusted source of how to make AI coding products better and more efficient.","@Napalys：https://github.com/napalys","Napalys Klicius is a Software Engineer at GitHub building agentic systems. His career has taken him from model checking to low-level C++ drone systems and static analysis, and more recently to teaching agents how to inspect code without getting lost.","Explore how the GitHub Copilot agentic harness delivers strong results across multiple benchmarks and leading token efficiency, while maintaining flexibility to choose among more than 20 models.","Kick off work in VS Code or the CLI, finish it from your phone. Remote control for GitHub Copilot sessions is now generally available on github.com and GitHub Mobile.","Managing library updates can be tedious at times. Learn how the GitHub Copilot app can handle this type of repetitive task.","These are the lessons we learned evaluating LLMs for real-world secret scanning.","We built a plugin for the GitHub Accessibility Scanner to make sure your alt text is actually accessible. Here’s how it works.","Everything you need to master GitHub, all in one place.","Build what’s next on GitHub, the place for anyone from anywhere to build anything.","Meet the companies and engineering teams that build with GitHub.","Join us October 28-29 in San Francisco or online for GitHub Universe, our flagship developer event uniting people, agents, and the world’s code."],"articleImages":[{"sourceUrl":"https://avatars.githubusercontent.com/u/54990334?v=4&s=200","alt":"Erik Kristensen","afterParagraph":0,"url":"/media/articles/cmtkfkfvl03garobqyr0vwpth/ac0d59d416e9f9f6.jpg"},{"sourceUrl":"https://avatars.githubusercontent.com/u/11835209?v=4&s=200","alt":"Napalys Klicius","afterParagraph":2,"url":"/media/articles/cmtkfkfvl03garobqyr0vwpth/36215561b44bb2d7.jpg"},{"sourceUrl":"https://github.blog/wp-content/uploads/2026/08/Screenshot-2026-08-26-at-12.57.57-PM.png?zoom=3&resize=400%2C212","alt":"GitHub Copilot app for Beginners: Work that runs itself.","afterParagraph":6,"url":"/media/articles/cmtkfkfvl03garobqyr0vwpth/85ff4984f98a59c0.webp"},{"sourceUrl":"https://github.blog/wp-content/uploads/2026/01/4ba0cd42388a255e04c78e5143548f22e577d68e0f15f68e6a3c76c18b927981-1920x1080-1.png?zoom=3&resize=400%2C212","alt":"Decorative background featuring floating green cubes, including one with the GitHub invertocat logo.","afterParagraph":7,"url":"/media/articles/cmtkfkfvl03garobqyr0vwpth/20fb41d9a2ff9940.webp"},{"sourceUrl":"https://github.blog/wp-content/uploads/2026/08/header-1.png?zoom=3&resize=400%2C212","alt":"Humorous header image with 'alt=IMG_2847.png' written across the top, and Mona's head looking up at a cube next to a GitHub logo on the bottom.","afterParagraph":8,"url":"/media/articles/cmtkfkfvl03garobqyr0vwpth/de2da5eb96831271.webp"}],"mediaStatus":"ok","articleBodyZh":["通过 GitHub Copilot CLI 将自主工作流程带到你的终端 >：https://github.com/features/copilot/cli","@erik-krogh：https://github.com/erik-krogh","Erik Krogh Kristensen 是 GitHub 的高级软件工程师，专注于人工智能、编码和安全的交叉领域。在 GitHub 的过去几年里，他参与了多种 AI 产品的开发，如 Copilot 自动修复、Copilot 代码审查，以及最近的 Copilot CLI。他是提高 AI 编码产品效率和性能的可信来源。","@Napalys：https://github.com/napalys","Napalys Klicius 是 GitHub 的软件工程师，专注于构建自主系统。他的职业生涯涵盖从模型检查到低级 C++ 无人机系统和静态分析，最近则专注于教导代理如何在检查代码时不迷路。","探索 GitHub Copilot 自主系统在多个基准测试中提供强大结果和领先的令牌效率，同时保持在 20 多种模型之间选择的灵活性。","在 VS Code 或 CLI 启动工作，从手机完成工作。GitHub Copilot 会话的远程控制现已在 github.com 和 GitHub Mobile 上普遍可用。","管理库更新有时可能很繁琐。了解 GitHub Copilot 应用如何处理这类重复任务。","这些是我们在评估用于现实世界秘密扫描的 LLM 时学到的经验教训。","我们为 GitHub 无障碍扫描器构建了一个插件，以确保你的替代文本真正可访问。以下是其工作原理。","掌握 GitHub 所需的一切，尽在一个地方。","在 GitHub 上构建下一步，无论你身处何地、是谁，都可以构建任何东西。","认识使用 GitHub 构建产品的公司和工程团队。","加入我们，10 月 28-29 日在旧金山或在线参加 GitHub Universe，这是我们的旗舰开发者活动，汇聚人类、代理和世界的代码。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"GitHub Blog介绍了Copilot在不牺牲任务质量的前提下降低AI编码成本的四项改动：选择性压缩工具输出、移除view工具行号前缀、压缩task-tool提示词，以及后台任务完成后直接交付结果。","background":"GitHub工程师Erik Kristensen分享，移除view工具行号前缀后，线下推理成本下降约5%，线上用户日均推理成本下降约3%；压缩task-tool提示词后，每轮约节省1300个token，归一化成本下降2.9%；后台任务结果直接交付后，AI Credits用量下降约2.3%。","viewpoint":"Aioga判断：这些措施集中在工具输出、提示词和任务交付流程，说明Copilot的成本优化并不只依赖模型选择，也涉及代理工作流中的信息组织与结果传递。","implications":"可能影响：若相关指标适用于对应测试和使用口径，工具输出与提示词压缩可能降低部分推理开销；但这些数据不代表所有任务、用户或模型场景都能获得相同幅度的变化，也不足以单独证明整体体验改善。","nextStep":"后续观察：需要关注GitHub是否披露各项改动对应的质量评估口径、适用范围和长期结果，并进一步区分线下推理成本、线上用户日均成本与AI Credits用量之间的统计边界。","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":"2026-09-02T19:06:57.866Z","sourceHash":"d75e664dd278c8ca","review":{"approved":true,"groundedness":93,"clarity":92,"duplicationRisk":18,"blockingIssues":[],"notes":["“Aioga判断”属于明确标注的分析性观点，未冒充来源事实。","“若相关指标适用于对应测试和使用口径”等限定语较为谨慎；可进一步将“说明”改为“表明其成本优化还涉及”以弱化因果确定性，但不影响审核通过。","建议核对“线下推理成本”“线上用户日均推理成本”和“AI Credits用量”的原文统计口径；候选内容已提示这些指标不可直接等同。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","editorial-labels","inference-boundary","low-source-overlap","no-html","independent-ai-review"]}},"tags":["行业动态","GitHub Blog"],"translations":{"zh-CN":{"title":"GitHub Copilot 如何在不牺牲任务质量的前提下降低 AI 编码成本","summary":"GitHub 工程师 Erik Kristensen 分享了 Copilot 降本的四项改动：选择性压缩工具输出、移除 view 工具行号前缀（线下推理成本降约 5%，线上用户日均推理成本降约 3%）、压缩 task-tool 提示词（每轮省约 1300 token，每活跃小时归一化成本降 2.9%）、后台任务完成后直接交付结果（AI Credits 用量降约 2.3%）。","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"GitHub Copilot 如何在不牺牲任务质量的前提下降低 AI 编码成本 - Aioga AI资讯","description":"GitHub 工程师 Erik Kristensen 分享了 Copilot 降本的四项改动：选择性压缩工具输出、移除 view 工具行号前缀（线下推理成本降约 5%，线上用户日均推理成本降约 3%）、压缩 task-tool 提示词（每轮省约 1300 token，每活跃小时归一化成本降 2.9%）、后台任务完成后直接交付结果（AI Credits 用量降...","url":"https://www.aioga.com/news/cmtkfkfvl03garobqyr0vwpth/","articleBody":["通过 GitHub Copilot CLI 将自主工作流程带到你的终端 >：https://github.com/features/copilot/cli","@erik-krogh：https://github.com/erik-krogh","Erik Krogh Kristensen 是 GitHub 的高级软件工程师，专注于人工智能、编码和安全的交叉领域。在 GitHub 的过去几年里，他参与了多种 AI 产品的开发，如 Copilot 自动修复、Copilot 代码审查，以及最近的 Copilot CLI。他是提高 AI 编码产品效率和性能的可信来源。","@Napalys：https://github.com/napalys","Napalys Klicius 是 GitHub 的软件工程师，专注于构建自主系统。他的职业生涯涵盖从模型检查到低级 C++ 无人机系统和静态分析，最近则专注于教导代理如何在检查代码时不迷路。","探索 GitHub Copilot 自主系统在多个基准测试中提供强大结果和领先的令牌效率，同时保持在 20 多种模型之间选择的灵活性。","在 VS Code 或 CLI 启动工作，从手机完成工作。GitHub Copilot 会话的远程控制现已在 github.com 和 GitHub Mobile 上普遍可用。","管理库更新有时可能很繁琐。了解 GitHub Copilot 应用如何处理这类重复任务。","这些是我们在评估用于现实世界秘密扫描的 LLM 时学到的经验教训。","我们为 GitHub 无障碍扫描器构建了一个插件，以确保你的替代文本真正可访问。以下是其工作原理。","掌握 GitHub 所需的一切，尽在一个地方。","在 GitHub 上构建下一步，无论你身处何地、是谁，都可以构建任何东西。","认识使用 GitHub 构建产品的公司和工程团队。","加入我们，10 月 28-29 日在旧金山或在线参加 GitHub Universe，这是我们的旗舰开发者活动，汇聚人类、代理和世界的代码。"]},"en":{"title":"How GitHub Copilot Reduces AI Coding Costs Without Sacrificing Task Quality","summary":"GitHub engineer Erik Kristensen shared four cost-reduction changes in Copilot: selectively compressing tool outputs, removing the line number prefix from the view tool (reduces offline inference cost by about 5% and the average daily online user inference cost by about 3%), compressing task-tool prompts (saves about 1300 tokens per round, normalized cost per active hour drops 2.9%), and delivering results directly after background tasks complete (AI Credits usage reduced by about 2.3%).","category":"Industry","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"How GitHub Copilot Reduces AI Coding Costs Without Sacrificing Task Quality - Aioga AI News","description":"GitHub engineer Erik Kristensen shared four cost-reduction changes in Copilot: selectively compressing tool outputs, removing the line number prefix from the view tool (reduces off...","url":"https://www.aioga.com/en/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:15.681Z"},"ja":{"title":"GitHub Copilotがタスクの質を犠牲にせずにAIコーディングコストを削減する方法","summary":"GitHubのエンジニア、エリック・クリステンセンはCopilotのコスト削減に向けた4つの変更点を共有しました:選択的圧縮ツール出力、ビューツールライン番号プレフィックスの削除(オフライン推論コスト約5%減少、オンラインユーザーの平均1日推論コスト約3%減少)、タスクツールプロンプト圧縮(1ラウンドあたり約1,300トークン節約、アクティブ時間あたりの正規化コスト2.9%削減)、そしてバックエンドタスク完了直後に結果を提供すること(AIクレジット使用率約2.3%減少)。","category":"業界動向","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"GitHub Copilotがタスクの質を犠牲にせずにAIコーディングコストを削減する方法 - Aioga AIニュース","description":"GitHubのエンジニア、エリック・クリステンセンはCopilotのコスト削減に向けた4つの変更点を共有しました:選択的圧縮ツール出力、ビューツールライン番号プレフィックスの削除(オフライン推論コスト約5%減少、オンラインユーザーの平均1日推論コスト約3%減少)、タスクツールプロンプト圧縮(1ラウンドあたり約1,300トークン節約、アクティブ時間あたりの正規...","url":"https://www.aioga.com/ja/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:18.138Z"},"ko":{"title":"GitHub Copilot이 작업 품질을 희생하지 않고 AI 코딩 비용을 낮추는 방법","summary":"GitHub 엔지니어 Erik Kristensen는 Copilot 비용 절감의 네 가지 변경 사항을 공유했습니다: 도구 출력 선택적 압축, view 도구 줄 번호 접두사 제거(오프라인 추론 비용 약 5% 감소, 온라인 사용자의 일일 평균 추론 비용 약 3% 감소), task-tool 프롬프트 압축(매 회 약 1300 토큰 절약, 활성 시간 기준 정규화 비용 2.9% 감소), 백그라운드 작업 완료 후 결과 직접 제공(AI 크레딧 사용량 약 2.3% 감소).","category":"업계 동향","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"GitHub Copilot이 작업 품질을 희생하지 않고 AI 코딩 비용을 낮추는 방법 - Aioga AI 뉴스","description":"GitHub 엔지니어 Erik Kristensen는 Copilot 비용 절감의 네 가지 변경 사항을 공유했습니다: 도구 출력 선택적 압축, view 도구 줄 번호 접두사 제거(오프라인 추론 비용 약 5% 감소, 온라인 사용자의 일일 평균 추론 비용 약 3% 감소), task-tool 프롬프트 압축(매 회 약 1300 토큰...","url":"https://www.aioga.com/ko/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:25.629Z"},"es":{"title":"Cómo GitHub Copilot puede reducir los costes de programación de IA sin sacrificar la calidad de las tareas","summary":"El ingeniero de GitHub, Erik Kristensen, compartió cuatro cambios que ahorran costes para Copilot: salida de la herramienta de compresión selectiva, eliminación del prefijo de número de línea de la herramienta de visualización (coste de inferencia offline reducido aproximadamente un 5%, coste medio diario de inferencia para usuarios online alrededor del 3%), compresión de prompts de herramientas de tarea (ahorro de unos 1.300 tokens por ronda, coste normalizado del 2,9% por hora activa) y entrega de resultados justo después de completar las tareas de backend (el uso de créditos de IA disminuyó alrededor del 2,3%).","category":"Industria","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Cómo GitHub Copilot puede reducir los costes de programación de IA sin sacrificar la calidad de las tareas - Aioga Noticias de IA","description":"El ingeniero de GitHub, Erik Kristensen, compartió cuatro cambios que ahorran costes para Copilot: salida de la herramienta de compresión selectiva, eliminación del prefijo de núme...","url":"https://www.aioga.com/es/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:26.959Z"},"fr":{"title":"Comment GitHub Copilot peut réduire les coûts de codage IA sans sacrifier la qualité des tâches","summary":"L’ingénieur de GitHub Erik Kristensen a partagé quatre changements permettant d’économiser pour Copilot : la sortie de l’outil de compression sélective, la suppression du préfixe du numéro de ligne d’outil de vue (coût d’inférence hors ligne en baisse d’environ 5 %, coût quotidien moyen d’inférence pour les utilisateurs en ligne en baisse d’environ 3 %), compression des invites de l’outil de tâche (économie d’environ 1 300 jetons par tour, coût normalisé en baisse de 2,9 % par heure active), et la livraison des résultats directement après la réalisation des tâches backend (utilisation des crédits IA en baisse d’environ 2,3 %).","category":"Industrie","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Comment GitHub Copilot peut réduire les coûts de codage IA sans sacrifier la qualité des tâches - Aioga Actualités IA","description":"L’ingénieur de GitHub Erik Kristensen a partagé quatre changements permettant d’économiser pour Copilot : la sortie de l’outil de compression sélective, la suppression du préfixe d...","url":"https://www.aioga.com/fr/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:36.232Z"},"de":{"title":"Wie GitHub Copilot die KI-Codierungskosten senkt, ohne die Aufgabenqualität zu beeinträchtigen","summary":"GitHub-Ingenieur Erik Kristensen teilte vier Änderungen zur Kostensenkung von Copilot: selektive Komprimierung der Tool-Ausgabe, Entfernung des Zeilennummern-Präfixes im View-Tool (Offline-Inferenzkosten um ca. 5% gesenkt, Online-Durchschnittskosten pro Benutzer um ca. 3%), Komprimierung der Task-Tool-Prompts (jede Runde spart ca. 1300 Token, normalisierte Kosten pro aktiver Stunde um 2,9% gesenkt), direkte Lieferung der Ergebnisse nach Abschluss von Hintergrundaufgaben (AI-Credits-Nutzung um ca. 2,3% gesenkt).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Wie GitHub Copilot die KI-Codierungskosten senkt, ohne die Aufgabenqualität zu beeinträchtigen - Aioga KI-News","description":"GitHub-Ingenieur Erik Kristensen teilte vier Änderungen zur Kostensenkung von Copilot: selektive Komprimierung der Tool-Ausgabe, Entfernung des Zeilennummern-Präfixes im View-Tool...","url":"https://www.aioga.com/de/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:33.363Z"},"pt-BR":{"title":"Como o GitHub Copilot reduz o custo de codificação por IA sem sacrificar a qualidade das tarefas","summary":"O engenheiro do GitHub, Erik Kristensen, compartilhou quatro mudanças do Copilot para reduzir custos: compressão seletiva da saída das ferramentas, remoção do prefixo de número de linha da ferramenta view (reduzindo o custo de inferência offline em cerca de 5% e o custo de inferência diário do usuário online em cerca de 3%), compressão do prompt task-tool (economizando cerca de 1300 tokens por rodada, reduzindo o custo normalizado por hora ativa em 2,9%) e entrega direta do resultado após a conclusão das tarefas em segundo plano (reduzindo o uso de Créditos de IA em cerca de 2,3%).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Como o GitHub Copilot reduz o custo de codificação por IA sem sacrificar a qualidade das tarefas - Aioga Notícias de IA","description":"O engenheiro do GitHub, Erik Kristensen, compartilhou quatro mudanças do Copilot para reduzir custos: compressão seletiva da saída das ferramentas, remoção do prefixo de número de...","url":"https://www.aioga.com/pt-BR/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:43.455Z"},"ru":{"title":"Как GitHub Copilot может снизить затраты на программирование ИИ, не жертвуя качеством задач","summary":"Инженер GitHub Эрик Кристенсен поделился четырьмя изменениями, экономящими средства для Copilot: выборочный вывод инструмента сжатия, удаление префикса номера строки инструмента (снижение стоимости офлайн-инференции примерно на 5%, средняя ежедневная стоимость вывода для пользователей онлайн снизилась примерно на 3%), сжатие подсказок для инструментов задач (экономия около 1 300 токенов за раунд, нормализованная стоимость снижение на 2,9% за активный час) и получение результатов сразу после завершения бэкенд-задач (использование AI Credits снизилось примерно на 2,3%).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Как GitHub Copilot может снизить затраты на программирование ИИ, не жертвуя качеством задач - Aioga Новости ИИ","description":"Инженер GitHub Эрик Кристенсен поделился четырьмя изменениями, экономящими средства для Copilot: выборочный вывод инструмента сжатия, удаление префикса номера строки инструмента (с...","url":"https://www.aioga.com/ru/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:45.081Z"},"ar":{"title":"كيفية خفض تكلفة الترميز باستخدام GitHub Copilot دون التضحية بجودة المهام","summary":"شارك مهندس GitHub إيريك كريستنسن أربع تغييرات لتقليل تكلفة Copilot: ضغط إخراج الأداة بشكل انتقائي، إزالة بادئة أرقام أسطر أداة view (خفض تكلفة الاستدلال غير المتصل بالإنترنت حوالي 5٪، وخفض تكلفة الاستدلال اليومي للمستخدمين عبر الإنترنت حوالي 3٪)، ضغط كلمات التلميح task-tool (توفير حوالي 1300 توكن لكل جولة، وانخفاض التكلفة الموحدة لكل ساعة نشطة بنسبة 2.9٪)، تسليم النتائج مباشرة بعد اكتمال المهام في الخلفية (انخفاض استخدام رصيد الذكاء الاصطناعي حوالي 2.3٪).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"كيفية خفض تكلفة الترميز باستخدام GitHub Copilot دون التضحية بجودة المهام - Aioga أخبار الذكاء الاصطناعي","description":"شارك مهندس GitHub إيريك كريستنسن أربع تغييرات لتقليل تكلفة Copilot: ضغط إخراج الأداة بشكل انتقائي، إزالة بادئة أرقام أسطر أداة view (خفض تكلفة الاستدلال غير المتصل بالإنترنت حوالي...","url":"https://www.aioga.com/ar/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:51.509Z"},"hi":{"title":"GitHub Copilot बिना कार्य की गुणवत्ता को बलिदान किए AI कोडिंग लागत कैसे कम करता है","summary":"GitHub इंजीनियर Erik Kristensen ने Copilot की लागत कम करने के चार बदलाव साझा किए: टूल आउटपुट का चयनात्मक संपीड़न, view टूल लाइन नंबर प्रिफिक्स हटाना (ऑफ़लाइन इन्फ़्रेंस लागत लगभग 5% कम हुई, ऑनलाइन उपयोगकर्ता की दैनिक औसत इन्फ़्रेंस लागत लगभग 3% कम हुई), task-tool प्रॉम्प्ट वाक्यांशों को संपीड़ित करना (प्रत्येक राउंड लगभग 1300 टोकन बचाए, प्रत्येक सक्रिय घंटे के लिए सामान्यीकृत लागत 2.9% कम हुई), बैकग्राउंड टास्क पूरा होने के बाद सीधे परिणाम देना (AI Credits का उपयोग लगभग 2.3% कम हुआ)।","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"GitHub Copilot बिना कार्य की गुणवत्ता को बलिदान किए AI कोडिंग लागत कैसे कम करता है - Aioga AI समाचार","description":"GitHub इंजीनियर Erik Kristensen ने Copilot की लागत कम करने के चार बदलाव साझा किए: टूल आउटपुट का चयनात्मक संपीड़न, view टूल लाइन नंबर प्रिफिक्स हटाना (ऑफ़लाइन इन्फ़्रेंस लागत लगभग 5...","url":"https://www.aioga.com/hi/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:52.251Z"},"it":{"title":"Come GitHub Copilot può ridurre i costi di codifica dell'IA senza sacrificare la qualità dei compiti","summary":"L'ingegnere di GitHub Erik Kristensen ha condiviso quattro cambiamenti che permettono di risparmiare costi per Copilot: output dello strumento di compressione selettiva, rimozione del prefisso del numero di linea dello strumento di visualizzazione (costo di inferenza offline in calo di circa il 5%, costo medio giornaliero di inferenza per gli utenti online in calo di circa il 3%), compressione dei prompt per strumento di lavoro (risparmio di circa 1.300 token per round, costo normalizzato in calo del 2,9% per ora attiva) e consegna dei risultati immediatamente dopo il completamento dei task backend (uso di crediti AI in calo di circa il 2,3%).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Come GitHub Copilot può ridurre i costi di codifica dell'IA senza sacrificare la qualità dei compiti - Aioga Notizie IA","description":"L'ingegnere di GitHub Erik Kristensen ha condiviso quattro cambiamenti che permettono di risparmiare costi per Copilot: output dello strumento di compressione selettiva, rimozione...","url":"https://www.aioga.com/it/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:04:01.100Z"},"nl":{"title":"Hoe GitHub Copilot AI-coderingkosten kan verlagen zonder concessies te doen aan de kwaliteit van taken","summary":"GitHub-ingenieur Erik Kristensen deelde vier aanpassingen van Copilot om kosten te besparen: selectief comprimeren van tool-output, verwijderen van rijnummerprefix van view-tools (offline inferentiekosten dalen met ongeveer 5%, dagelijkse inferentiekosten voor online gebruikers dalen met ongeveer 3%), comprimeren van task-tool prompts (ongeveer 1300 tokens per ronde bespaard, genormaliseerde kosten per actieve uur dalen met 2,9%), resultaten rechtstreeks leveren na voltooiing van achtergrondtaken (gebruik van AI Credits daalt met ongeveer 2,3%).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Hoe GitHub Copilot AI-coderingkosten kan verlagen zonder concessies te doen aan de kwaliteit van taken - Aioga AI-nieuws","description":"GitHub-ingenieur Erik Kristensen deelde vier aanpassingen van Copilot om kosten te besparen: selectief comprimeren van tool-output, verwijderen van rijnummerprefix van view-tools (...","url":"https://www.aioga.com/nl/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:03:58.965Z"},"tr":{"title":"GitHub Copilot, görev kalitesinden ödün vermeden yapay zeka kodlama maliyetlerini nasıl azaltabilir?","summary":"GitHub mühendisi Erik Kristensen, Copilot için dört maliyet tasarrufu sağlayan değişiklik paylaştı: seçici sıkıştırma aracı çıktısı, görüntüleme aracı satırı numarası önekinin kaldırılması (çevrimdışı çıkarım maliyeti yaklaşık %5 düştü, çevrimiçi kullanıcılar için ortalama günlük çıkarım maliyeti yaklaşık %3 düştü), görev aracı promptu sıkıştırma (tur başına yaklaşık 1.300 token tasarrufu sağlamak, aktif saat başına %2,9 azalan normalize maliyet) ve arka uç görevleri tamamlandıktan hemen sonra sonuç vermek (AI kredisi kullanımı yaklaşık %2,3 azaldı).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"GitHub Copilot, görev kalitesinden ödün vermeden yapay zeka kodlama maliyetlerini nasıl azaltabilir? - Aioga AI Haberleri","description":"GitHub mühendisi Erik Kristensen, Copilot için dört maliyet tasarrufu sağlayan değişiklik paylaştı: seçici sıkıştırma aracı çıktısı, görüntüleme aracı satırı numarası önekinin kald...","url":"https://www.aioga.com/tr/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:04:09.995Z"},"vi":{"title":"GitHub Copilot làm thế nào để giảm chi phí lập trình AI mà không làm giảm chất lượng công việc","summary":"Kỹ sư GitHub Erik Kristensen chia sẻ bốn thay đổi giúp Copilot giảm chi phí: chọn lọc nén đầu ra công cụ, loại bỏ tiền tố số dòng của công cụ view (giảm khoảng 5% chi phí suy luận ngoại tuyến và khoảng 3% chi phí suy luận trung bình hàng ngày của người dùng trực tuyến), nén từ khóa task-tool (tiết kiệm khoảng 1300 token mỗi vòng, giảm 2.9% chi phí tính theo giờ hoạt động), và giao kết quả trực tiếp sau khi hoàn thành tác vụ nền (giảm khoảng 2.3% lượng AI Credits).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"GitHub Copilot làm thế nào để giảm chi phí lập trình AI mà không làm giảm chất lượng công việc - Tin tức AI Aioga","description":"Kỹ sư GitHub Erik Kristensen chia sẻ bốn thay đổi giúp Copilot giảm chi phí: chọn lọc nén đầu ra công cụ, loại bỏ tiền tố số dòng của công cụ view (giảm khoảng 5% chi phí suy luận...","url":"https://www.aioga.com/vi/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:04:07.963Z"},"id":{"title":"Bagaimana GitHub Copilot Mengurangi Biaya Kode AI Tanpa Mengorbankan Kualitas Tugas","summary":"Insinyur GitHub Erik Kristensen membagikan empat perubahan Copilot untuk mengurangi biaya: memilih untuk mengompresi output alat secara selektif, menghapus awalan nomor baris alat view (biaya inferensi offline turun sekitar 5%, biaya inferensi rata-rata pengguna online turun sekitar 3%), mengompresi prompt task-tool (menghemat sekitar 1300 token per putaran, biaya per jam aktif turun 2,9%), dan mengirim hasil langsung setelah tugas latar belakang selesai (penggunaan AI Credits turun sekitar 2,3%).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Bagaimana GitHub Copilot Mengurangi Biaya Kode AI Tanpa Mengorbankan Kualitas Tugas - Berita AI Aioga","description":"Insinyur GitHub Erik Kristensen membagikan empat perubahan Copilot untuk mengurangi biaya: memilih untuk mengompresi output alat secara selektif, menghapus awalan nomor baris alat...","url":"https://www.aioga.com/id/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:04:16.890Z"},"th":{"title":"วิธีที่ GitHub Copilot สามารถลดต้นทุนการเขียนโค้ด AI โดยไม่ลดคุณภาพของงาน","summary":"วิศวกร GitHub Erik Kristensen ได้แชร์การเปลี่ยนแปลงสี่ประการที่ช่วยประหยัดต้นทุนสําหรับ Copilot: การคัดเลือกผลลัพธ์เครื่องมือบีบอัด, การลบคํานําหน้าหมายเลขบรรทัดเครื่องมือ (ต้นทุนการอนุมานแบบออฟไลน์ลดลงประมาณ 5%, ต้นทุนการอนุมานเฉลี่ยรายวันสําหรับผู้ใช้ออนไลน์ลดลงประมาณ 3%), การบีบอัดพรอมต์ของเครื่องมือ (ประหยัดโทเค็นได้ประมาณ 1,300 โทเคนต่อรอบ, ลดต้นทุนที่ปรับให้เหมาะสม 2.9% ต่อชั่วโมงที่ใช้งาน), และการแสดงผลลัพธ์โดยตรงหลังจากงาน backend เสร็จสิ้น (การใช้เครดิต AI ลดลงประมาณ 2.3%)","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"วิธีที่ GitHub Copilot สามารถลดต้นทุนการเขียนโค้ด AI โดยไม่ลดคุณภาพของงาน - ข่าว AI Aioga","description":"วิศวกร GitHub Erik Kristensen ได้แชร์การเปลี่ยนแปลงสี่ประการที่ช่วยประหยัดต้นทุนสําหรับ Copilot: การคัดเลือกผลลัพธ์เครื่องมือบีบอัด, การลบคํานําหน้าหมายเลขบรรทัดเครื่องมือ (ต้นทุนก...","url":"https://www.aioga.com/th/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:04:19.644Z"},"pl":{"title":"Jak GitHub Copilot może obniżyć koszty kodowania AI bez poświęcania jakości zadań","summary":"Inżynier GitHub, Erik Kristensen, podzielił się czterema zmianami oszczędzającymi koszty dla Copilota: wynik narzędzia selektywnej kompresji, usunięcie prefiksu linii narzędzi widoku (koszt wnioskowania offline spadł o około 5%, średni dzienny koszt wnioskowania dla użytkowników online o około 3%), kompresja promptów narzędzi zadaniowych (oszczędzając około 1 300 tokenów na rundę, normalizowany koszt spadł o 2,9% za godzinę aktywną) oraz dostarczanie wyników bezpośrednio po zakończeniu zadań backendowych (zużycie kredytów AI spadło o około 2,3%).","category":"行业动态","source":"GitHub Blog","aggregationSource":"GitHub Blog","pageTitle":"Jak GitHub Copilot może obniżyć koszty kodowania AI bez poświęcania jakości zadań - Aioga Wiadomości AI","description":"Inżynier GitHub, Erik Kristensen, podzielił się czterema zmianami oszczędzającymi koszty dla Copilota: wynik narzędzia selektywnej kompresji, usunięcie prefiksu linii narzędzi wido...","url":"https://www.aioga.com/pl/news/cmtkfkfvl03garobqyr0vwpth/","contentTranslated":true,"sourceHash":"d26e5d4c3c41b23f","translatedAt":"2026-09-02T19:04:27.387Z"}},"evidenceTier":"verified-news","reviewStatus":"editorial-selected","indexable":true,"editorialCover":""}}