跳到正文
Claude Platform：开发者版本说明·· 8 天前AI 评分74

Claude Platform 版本说明：发布 Claude Haiku 5.5，Sonnet 5.5 缓存读取降价至 $0.10/百万 token

Claude Platform release notes — October 7, 2026

AI 导读

Anthropic 于 10 月 7 日发布 Claude Haiku 5.5（claude-haiku-5-5），面向高并发和低延迟场景，支持 1M token 上下文窗口、128k 最大输出 token 和带 effort 参数的自适应思考，已在 Claude API、Amazon Bedrock、Google Cloud 和 Microsoft Foundry 上线。

正文 · AI 翻译

译文尚不完整，完整内容请切换到原文。

发布说明

Claude 平台的更新，包括 Claude API、客户端 SDK 和 Claude Console。

Claude 平台发布说明列出了 Claude API、客户端 SDK 和 Claude Console 的变更，最新在前。

2026 年 10 月 8 日

2026 年 10 月 7 日

  • 我们将 Claude Sonnet 5.5 的提示缓存读取价格从每百万 token 0.20 美元降至 0.10 美元：即基础输入价格的 0.05 倍，而非 0.1 倍。缓存写入和所有其他价格保持不变。请参阅提示缓存定价。

2026 年 10 月 7 日

2026 年 10 月 7 日

2026 年 10 月 7 日

2026 年 10 月 7 日

  • 在 Claude Managed Agents 中，具有 limited 网络连接的云环境现在也会将其 allowed_hosts 应用于 web_search 和 web_fetch 工具。对于 allowed_hosts 不匹配的主机上的 URL，web_fetch 调用会向 agent 返回 url_not_allowed 错误结果。web_search 会省略来自此类主机的结果。当 allowed_hosts 未列出任何主机时，这两个工具都不会返回页面或搜索结果。allow_package_managers 和 allow_mcp_servers 不会为这些工具添加主机。要让这些工具访问某个主机，请将其添加到 allowed_hosts，这也会将其开放给沙箱。unrestricted 网络连接和自托管环境不会限制这些工具。请参阅环境网络连接。
  • 使用 limited 网络连接时，如果某个已启用的 web 工具的 allowed_domains 包含不在 allowed_hosts 内的条目，则创建会话会失败并返回 400 错误。添加此类条目的会话更新也会如此。allowed_hosts 条目会匹配一个精确主机，除非它以 *. 开头，因此 docs.example.com 不在 ["example.com"] 内。要修复该错误，请将该主机添加到 allowed_hosts，或从 allowed_domains 中移除该条目。请参阅限制 web 搜索和 web 抓取域名。

2026 年 10 月 6 日

  • 我们在 Models API 中新增了 capabilities.server_tools。GET /v1/models 和 GET /v1/models/{model_id} 现在会报告每个模型是否接受 web search 和 code execution 工具。要检查某个模型是否接受 code execution 工具，请读取 capabilities.server_tools.code_execution。顶层的 capabilities.code_execution 会报告代码是否可以调用你请求中的其他工具。参见 Using the Models API。

October 5, 2026

  • 我们在 Models API 中新增了 capabilities.thinking.types.disabled。GET /v1/models 和 GET /v1/models/{model_id} 现在会报告每个模型是否接受 thinking: {type: "disabled"}，该参数会关闭思考。参见 Using the Models API。

October 1, 2026

  • 我们在 Models API 中新增了 line 字段。GET /v1/models 和 GET /v1/models/{model_id} 现在会返回每个模型所属的模型系列。例如，Claude Opus 4.5 和 Claude Opus 4.6 都报告 opus。使用 line 即可对模型进行分组，无需解析其 ID。对于不属于任何系列的模型，line 为 null。参见 Using the Models API。

September 30, 2026

  • 在 Python SDK 1.10.0、TypeScript SDK 0.130.0、C# SDK 12.52.0、Go SDK 1.77.0、Java SDK 2.67.0、PHP SDK 0.53.0、Ruby SDK 1.75.0 以及 ant CLI 的 1.37.0 版本中，用于组织信息、成员、邀请、工作区、API 密钥、速率限制、服务账户、工作负载身份联合、客户托管加密密钥和合规设置的 Admin API 方法已结束 beta：在 client.organization 和 ant organization 下调用它们。这些版本还保留了 client.beta.organization 和 ant beta:organization。
  • 我们宣布弃用 Claude Sonnet 4.5 模型（claude-sonnet-4-5-20250929），其在 Claude API 上的退役时间定于 2026 年 11 月 30 日。我们建议迁移到 Claude Sonnet 5.5。更多内容请阅读 Model deprecations。

September 28, 2026

  • 我们推出了 Claude Sonnet 5.5（claude-sonnet-5-5）。它可在 Claude API、Claude in Amazon Bedrock、Claude Platform on AWS、Claude on Google Cloud 和 Claude in Microsoft Foundry 上使用。有关其上下文窗口、输出限制和价格，请参见 Claude Sonnet 5.5 model page。
  • 为 Claude Sonnet 5 编写的代码在 Claude Sonnet 5.5 上可能以五种方式出错。要关闭前置思考，请在 high 及以下 effort 时发送 thinking: {"type": "between_tools"} 而不是 "disabled"。强制工具使用（tool_choice 类型 any 和 tool）会返回 400 错误。思考块与模型和对话绑定。在 Claude API 和 Google Cloud 上，不再接受较早的 computer_20251124 computer use 工具。advisor 工具拒绝将 Claude Opus 4.8、Claude Opus 4.7 和 Claude Sonnet 5 作为 advisor。有关每项变更，请参见 What's new in Claude Sonnet 5.5；有关变更前后的请求，请参见 migration guide。有关特定于模型的提示模式，请参见 Prompting Claude Sonnet 5.5。
  • Claude Sonnet 5.5 生成的思考块仅在其生成的账户中或与该账户关联的账户中有效。当另一个账户发送这些块之一时，API 会在模型看到该块之前将其丢弃，请求会成功。较早模型的块不受影响。参见 Preserved thinking。

September 24, 2026

  • 当 stop_details.category 为 "bio"、"frontier_llm" 或 "reasoning_extraction" 时，对于在任何输出之前到达的拒绝，我们将恢复计费，这些类别中我们测得的误报量较低。流中途的拒绝此前已计费。根据此变更计费的拒绝，按运行它的模型的费率，像其他任何请求一样收费。其他类别中在任何输出之前的拒绝仍不计费，回退额度保持不变。此变更适用于所有平台。参见拒绝如何计费。
  • Compliance API 的本地会话端点已结束测试，适用于 Excel、PowerPoint、Word 和 Outlook 中的 Claude for Microsoft 365 会话（product_surface 值以 office_agents 开头）。参见用户机器上的会话。
  • Compliance API 的活动动态不再返回文件名、项目文档名称或工件标题。文件、项目文档和工件活动上的 filename 和 title 字段现在始终为空或被省略，包括在此变更之前记录的活动。要按活动上的 ID 查找名称或标题，请使用具有 read:compliance_user_data 范围的 Compliance Access Key。参见了解 Activity 对象。

2026 年 9 月 23 日

  • 缓存诊断已在 Claude API 上结束测试，不再需要 cache-diagnosis-2026-04-07 beta 标头。在 Messages 请求中包含 diagnostics 对象即可启用；仍发送该标头的请求照常工作。来自 POST /v1/messages 的响应现在始终包含 diagnostics 字段，当请求未包含 diagnostics 对象时该字段为 null。

2026 年 9 月 22 日

  • 我们推出了 Claude Opus 5.5（claude-opus-5-5），这是一款面向长时间运行的智能体编程和知识工作的模型。它默认具有100 万 token 上下文窗口、128k 最大输出 token，以及始终开启的自适应思考，价格为每百万 token 4 美元 / 20 美元（Claude Opus 5 为 5 美元 / 25 美元）。Claude Opus 5.5 可在 Claude API、Amazon Bedrock 中的 Claude、AWS 上的 Claude Platform、Google Cloud 上的 Claude 以及 Microsoft Foundry 中的 Claude 上使用。有关功能、API 变更和迁移指南，请参见Claude Opus 5.5 的新增内容。
  • 在 Claude Opus 5.5 上，思考无法禁用：thinking: {"type": "disabled"} 和 thinking: {"type": "enabled", ...} 会返回 400 错误。省略 thinking 字段，并使用effort 参数控制思考深度。tool_choice 类型 any 和 tool 也会返回 400 错误，与 Claude Fable 5.1 上相同；请将 auto 与严格工具使用配合使用。在 Claude API 和 Google Cloud 上，此模型上的计算机使用需要 computer_toolset_20260801 工具集，而较早的 computer_20251124 工具会返回 400 错误；在 Amazon Bedrock 上，computer_20251124 继续可用。参见迁移指南。
  • 快速模式（研究预览版）已在 Claude API 上适用于 Claude Opus 5.5。
  • 现在可以在对话中途的系统消息中定义工具，该功能在 Claude API 上以 beta 形式提供，需使用 inline-tools-2026-09-15 beta 标头。一个 tool_addition 块可以携带工具的完整定义（tool: {"type": "tool_definition", "definition": {...}}），因此你可以添加工具、更改其 schema，或将服务器工具迁移到更新版本，而无需编辑 tools 或使提示缓存失效。同一标头也涵盖通过引用添加和移除工具。再加上 MCP 连接器的 mcp-client-2026-09-15 beta 标头，定义可以是 MCP 工具集，响应会在 mcp_tool_listing 块中记录每个服务器获取的工具列表，当你将其发回时该列表会被固定。

2026 年 9 月 18 日

  • 对于缓存诊断，发送 cache-diagnosis-2026-04-07 beta 标头的请求，其响应现在始终包含 diagnostics 字段。当请求未包含 null 对象时，该字段为 diagnostics。此前在这种情况下该字段会被省略。
  • Compliance API 的本地会话端点现在还会返回 Claude in Chrome 会话的记录（product_surface 值为 claude_in_chrome），在 Claude Enterprise 组织中为 beta 版，使用您现有的 Compliance Access Key 和 read:compliance_user_data 作用域。请参阅用户机器上的会话。

2026 年 9 月 14 日

  • Messages API 现在可以在 Claude API 上按需压缩对话，使用 compact-2026-09-04 beta 标头时为 beta 版。发送顶层 compaction 参数，API 会返回一个带签名的 compaction 块，用于汇总您发送的消息。在后续请求中，先发送该块，以替代那些消息。您可以自行选择何时压缩，请求可以在后台运行，并且您可以在摘要之后逐字保留最近的轮次。在支持保留思考的模型上，这些保留轮次中的思考可以保持有效。
  • 使用 thinking-binding-controls-2026-08-01 beta 标头时，input_transformations 响应字段会新增第二种条目类型 thinking_mismatch_allowed。它表示一个思考块，该块在 API 不强制执行前缀检查的请求中未通过前缀检查：例如，在 Claude Fable 5.1 上，来自 2026 年 8 月 31 日之前创建的账户且未设置 prefix_mismatch_behavior 的请求。该块仍会原样到达模型。在您选择启用强制执行之前，记录这些条目以发现生产流量中的历史编辑。请参阅设置不匹配行为并读取 input_transformations。

2026 年 9 月 10 日

  • Claude Managed Agents 权限策略现在包含 auto：服务器会评估每个 agent 或 MCP 工具调用，并运行、拒绝或暂停以等待您的批准。agent.tool_use 和 agent.mcp_tool_use 事件会在 evaluation 字段中报告每个调用是如何被评估的，同时附带 evaluated_permission。请参阅让服务器使用 auto 评估每个调用。
  • ant CLI 的 1.32.0 版本新增了 ant beta:sessions connect，可将您的终端连接到 Claude Managed Agents 会话。您可以实时跟踪会话、发送消息，并允许或拒绝正在等待批准的工具调用。传入 --web 可在本地提供 Claude Console 的会话查看器，并在那里打开会话。请参阅从终端连接到 Managed Agents 会话。

2026 年 9 月 9 日

  • 对于缓存诊断，API 现在仅在请求包含 diagnostics 对象时才存储请求的指纹。仅发送 cache-diagnosis-2026-04-07 beta 标头的请求仍会被接受，但不会存储指纹。后续轮次若将 previous_message_id 指向它，会报告 previous_message_not_found。请在每一轮都包含 diagnostics，并在第一轮包含 "previous_message_id": null。

2026 年 9 月 3 日

  • ant CLI 的 1.30.0 版本新增了 ant apply，可从您仓库中的文件创建和更新 agents、environments、skills、memory stores 和 deployments。在文件中描述每个资源，运行 ant apply，并批准它打印的计划。提交它写入的 claude-lock.json 锁文件，以便后续运行（在您的机器上或 CI 中）更新相同的资源，而不是创建新资源。请参阅使用 ant apply 以代码方式管理资源。
  • 按消息努力的更改（beta 版）也已在 Google Cloud 上适用于 Claude Fable 5.1、Claude Mythos 5.1 和 Claude Opus 5，使用相同的 mid-conversation-output-config-2026-07-01 beta 标头。

2026 年 9 月 1 日

  • 我们推出了 Claude Fable 5.1（claude-fable-5-1），它是 Claude Fable 5 的继任者，面向长时间运行的智能体编程、知识工作和研究，同时推出了面向 Project Glasswing 参与者的 Claude Mythos 5.1（claude-mythos-5-1）。两个模型默认支持 1M token 上下文窗口、128k 最大输出 token，以及始终开启的 自适应思考，价格为每 MTok $10 / $50 USD，与 Claude Fable 5 相同，缓存读取降至每 MTok $0.25。Claude Fable 5.1 可在 Claude API、Claude in Amazon Bedrock、Claude Platform on AWS、Claude on Google Cloud 和 Claude in Microsoft Foundry 上使用。有关功能、API 变更和迁移指南，请参阅 What's new in Claude Fable 5.1。
  • Claude Fable 5.1 和 Claude Mythos 5.1 上的提示缓存读取费用为每百万 token $0.25 USD：是基础输入价格的 0.025 倍，而其他模型为 0.1 倍。缓存写入保持不变。请参阅 Prompt caching pricing。
  • 在 Claude Fable 5.1 和 Claude Mythos 5.1 上，tool_choice 类型 any 和 tool 不受支持，并会返回 400 错误。auto 和 none 保持不变。为保证工具输入符合 schema，请使用 strict tool use 或 structured outputs。
  • Claude Fable 5.1 和 Claude Mythos 5.1 生成的思考块仅对生成它们的模型或更新的模型保留：更早的模型无法读取它们，API 会丢弃重放到更早模型的思考块。Claude Fable 5.1 接受来自 Claude Opus 5、Claude Fable 5、Claude Mythos 5 及更早 Claude 模型的思考块。在 Claude Fable 5.1 上，API 还会检查块之前的内容是否发生变化：对于 2026 年 8 月 31 日或之后创建的新账户，在 system 提示、tools 或更早消息发生变化后重放思考块会返回 400 错误。使用 thinking-binding-controls-2026-08-01 beta 标头时，被丢弃的块会在 input_transformations 响应字段中报告，thinking.block_binding.prefix_mismatch_behavior 可选择拒绝还是丢弃历史已变化的块。请参阅 Preserved thinking。
  • 按消息调整 effort 功能在 Claude API 上的 Claude Fable 5.1、Claude Mythos 5.1 和 Claude Opus 5 中处于 beta 阶段。在 messages 内添加一条包含 output_config.effort 的 role: "system" 消息，即可在保留提示缓存的同时更改后续轮次的 effort。请在请求中包含 mid-conversation-output-config-2026-07-01 beta 标头。请参阅 Per-message effort。
  • 轮次作用域系统消息处于 beta 阶段（mid-conversation-system-clear-at-2026-08-21 标头）。在对话中途的 role: "system" 消息上设置 clear_at: "next_user_message"，它仅对当前轮次生效，之后会以零 token 成本保留在历史记录中。每轮提醒不会累积，也不会使提示缓存或后续思考块失效。
  • thinking.display 在 beta 阶段（thinking-display-updates-2026-08-18 标头）接受第三个值 "updates"。推理返回时 thinking 字段为空，与 "omitted" 下相同，而 Claude Fable 5.1、Claude Mythos 5.1 和 Claude Fable 5 在工具调用之间写入的简短进度更新会以文本形式返回，在工具调用前最多有一个 thinking 块。请参阅 Progress updates between tool calls。
  • Claude Fable 5.1 和 Claude Mythos 5.1 生成的文本带有 Anthropic 的文本水印，而 Claude 通过代码执行工具生成的受支持图像、视频和音频文件，在你通过 Claude API 上的 Files API 检索时会带有 C2PA Content Credentials。标记无需对你的请求或响应处理做任何更改。
  • 与 Claude Fable 5 一样，这两个模型都要求 30 天数据保留，并且除非经 Anthropic 明确授权，否则在零数据保留下不可用。请参阅特定于模型的数据保留要求。
  • Admin API 的 Claude Enterprise 端点指南（用户管理和支出限额）、Claude Enterprise Analytics API 以及Compliance API 现在会显示 anthropic-version 标头；请像 Claude API 的其他部分一样，在每次请求这些端点时都发送它。请参阅API 版本。

2026 年 8 月 27 日

  • 在 Python SDK 1.2.0、TypeScript SDK 0.122.0、Go SDK 1.68.0、Java SDK 2.59.0、Ruby SDK 1.67.0 和 C# SDK 12.44.0 中，client.beta.files 和 client.beta.skills 不再发送 files-api-2025-04-14 和 skills-2025-10-02 beta 标头，并返回与 client.files 和 client.skills 相同的结构。通过此更改，client.beta.skills.delete() 会删除一个 Skill 及其所有版本，并且 beta Messages 类型 BetaSkill（容器 Skill 引用）被重命名为 BetaContainerSkill。仍然发送 beta 标头的请求会继续收到 beta 结构。请参阅从 files-api-2025-04-14 迁移和从 skills-2025-10-02 迁移。
  • 你现在可以在 Claude Console 中创建个人密钥和服务账户密钥。它们以你或服务账户的身份行事，具有相同的权限，并在关联账户从组织中移除后停止工作。这使组织管理员能够更轻松地跟踪每个账户的使用情况，并确保密钥使用是合法的。这些 API 密钥可以限定到特定工作区，或在管理端点上工作并跨该账户有权访问的任何工作区使用。工作区 API 密钥仍作为旧版选项受到支持。有关更多信息，请参阅API 密钥。

2026 年 8 月 26 日

  • Compliance API 会话端点已结束 beta，适用于 Cowork 和 Claude Code 会话。请参阅检索会话记录。
  • Compliance API 本地会话端点现在还返回 Claude Science 会话（product_surface 值 claude_science）以及 Excel、PowerPoint、Word 和 Outlook 中的 Claude for Microsoft 365 会话（product_surface 值以 office_agents 开头）的记录，在 Claude Enterprise 组织中为 beta，使用你现有的 Compliance Access Key 和 read:compliance_user_data 范围。请参阅用户机器上的会话。
  • Admin API 现已在 ant CLI 以及 Python、TypeScript、C#、Go、Java、PHP 和 Ruby SDK 中提供，位于 client.beta.organization 下。它们涵盖组织信息、成员、邀请、工作区和工作区成员、API 密钥、速率限制、服务账户、工作负载身份联合颁发者和规则，以及客户管理的加密密钥。使用情况和成本报告以及 Claude Enterprise 用户管理和分析端点仍仅支持 curl。CLI 和 SDK 从 ANTHROPIC_API_KEY 读取 Admin API 密钥，或从 ANTHROPIC_AUTH_TOKEN 读取 org:admin OAuth 令牌。

2026 年 8 月 20 日

  • 我们发布了 Python SDK v1.0。该 SDK 的 HTTP 层从 httpx 迁移到 httpx2，这是一个仍在维护、API 兼容的分支：可从 httpx2 构建自定义的 http_client、Timeout 和传输对象（DefaultHttpxClient 辅助函数保持不变），如果你依赖会修补 httpx 的追踪或模拟库，请在启动时调用 httpx2.alias_httpx()。v1.0 要求 Python 3.10 或更高版本，并移除了长期弃用的接口，包括旧版 Text Completions API、Messages 方法上的 temperature、top_p 和 top_k 参数，以及工具运行器的客户端 compaction_control。在异步客户端上，.with_raw_response 结果现在需要 await response.parse()，并且当未配置 AWS 区域时，AnthropicBedrock 现在会抛出错误，而不是默认使用 us-east-1。有关每一项变更及前后对比代码片段，请参阅 v1 迁移指南。
  • computer use 和 browser use 工具集（computer_toolset_20260801 和 browser_toolset_20260801）现已在 Google Cloud 上面向 Claude Fable 5、Claude Mythos 5、Claude Opus 5、Claude Sonnet 5 和 Claude Opus 4.8 提供。请求使用与 Claude API 上相同的 tools 条目。

2026年8月19日

  • computer use 工具已在 Claude API 上结束测试，成为 computer_toolset_20260801 工具集：无需 beta 标头、支持批量操作（一个回合内执行多个操作）、默认启用 zoom，并可通过 configs 进行按成员配置。更早的 beta 版本仍然可用。升级现有集成会改变请求结构和工具处理方式；请参阅 从 computer_20251124 迁移。
  • 我们推出了 browser use 工具（browser_toolset_20260801），这是一个客户端工具集，用于驱动由你的应用程序托管的浏览器。它在浏览器视口内运行，而不是整个桌面，直接读取页面本身（其无障碍树、元素、表单和标签页），并在截图加点击控制的基础上增加了元素引用、表单输入、标签页管理、下载报告和可选的文件上传功能。
  • 这两个工具集均可在 Claude API 上用于 Claude Fable 5、Claude Mythos 5、Claude Opus 5、Claude Sonnet 5 和 Claude Opus 4.8。
  • Files API 已在 Claude API 上结束测试。对 /v1/files 端点的请求，以及引用已上传文件的 Messages API 请求，不再需要 files-api-2025-04-14 beta 标头。不带该标头发送的请求使用当前响应格式：文件过期（上传文件时设置 expires_in_seconds；文件对象会报告 expires_at），以及 分页 page 和 next_page，外加在列出文件时的 ids[] 过滤器。仍然发送 beta 标头的 /v1/files 请求可继续正常工作，并返回之前的响应格式。 要将现有集成从该标头迁移出去，请参阅 从 files-api-2025-04-14 迁移。
  • Agent Skills 和 Skills API（/v1/skills）已在 Claude API 上结束测试。请求不再需要 skills-2025-10-02 beta 标头，包括通过 container 参数加载 Skills 的 Messages API 请求。仍然发送该标头的请求可继续正常工作，无需更改。请参阅 在 API 中使用 Agent Skills。 要将现有集成从该标头迁移出去，请参阅 从 skills-2025-10-02 迁移。
  • Claude Enterprise（claude.ai）组织的 Admin API 用户管理端点（成员、邀请、群组和自定义角色）已结束测试。群组和自定义角色请求不再需要 anthropic-beta: ce-user-management-2026-07-13 标头；仍然发送该标头的请求会被原样接受。请参阅 用户管理。
  • 现在你可以限制 Claude Managed Agents 智能体的 web_search 和 web_fetch 工具可以访问哪些站点。在 agent_toolset_20260401 configs 数组中该工具的条目上设置 allowed_domains 或 blocked_domains；web_fetch 也接受 max_content_tokens，web_search 接受 user_location。每个 configs 条目由其 name 标识，并由可选的 type 指定类型，仅传入 name、enabled 和 permission_policy 的请求仍可正常工作；在类型化 SDK 中，configs 条目会变为 按工具区分的类型。参见 限制网络搜索和网络抓取域名。
  • 在 自托管沙箱中运行的 Claude Managed Agents 会话现在可以挂载 记忆存储。Python、TypeScript 和 Go SDK worker 会将每个挂载的存储下载到沙箱中其 mount_path 处，并将智能体的更改同步回该存储。参见 自托管沙箱中的记忆存储。
  • Claude Console 中的会话查看器已重新设计，新增了时间线缩略图、按模型请求分组的对话记录，以及一个 Inspector 面板，用于显示会话详情和成本、原始事件、按工具统计、已挂载资源以及按线程的活动。参见 Console 可观测性。

2026 年 8 月 18 日

  • Claude Console 中的 Workbench 现在更名为 playground。Playground 支持所有 Messages API 参数，并包含演示代码执行和网络搜索等 API 功能的模板。它会显示每次运行的完整 SDK 请求和 API 响应，帮助你理解该 API 并基于它进行构建。更多信息请参见 Claude 帮助中心，或在 platform.claude.com/playground 试用。

2026 年 8 月 11 日

  • Compliance API 现在会返回在用户机器上运行的 Cowork 和 Claude Code 会话的对话记录，面向 Claude Enterprise 组织提供 beta 版。GET /v1/compliance/apps/sessions/local 列出你组织内的会话，GET /v1/compliance/apps/sessions/local/{session_id} 检索单个会话的元数据，GET /v1/compliance/apps/sessions/local/{session_id}/messages 返回其对话记录，全部使用你现有的 Compliance Access Key 和 read:compliance_user_data 权限范围。参见 用户机器上的会话。
  • 我们为 Claude API 添加了 anthropic-workspace-id 响应头。它携带请求的 API 密钥或访问令牌所解析到的工作区的 wrkspc_ 前缀 ID，包括你组织的 Default Workspace。参见 识别 API 响应背后的工作区。

2026 年 8 月 10 日

  • Claude Sonnet 5 的 introductory pricing（每百万 token 2 美元 / 10 美元）现已成为标准价格：原定于 2026 年 9 月 1 日上调至每百万 token 3 美元 / 15 美元的计划将不再执行。参见 定价。

2026 年 8 月 7 日

  • 你现在可以为 Claude Managed Agents 会话设置预算：即该会话支出的硬性上限，按公开标价计算。达到预算的会话会以 budget_reached 停止原因暂停，而不会发起新的模型请求；更改或移除预算即可恢复。Deployments 接受相同的预算，并将其应用于它们启动的每个会话。参见 会话预算。
  • 你现在可以为 Claude Managed Agents 会话指定一位顾问：一个能力至少与智能体自身相当的模型，会话的主线程可以在回合中途咨询它以获取战略指导。将其配置为智能体多智能体名册中的 {"type": "advisor"} 条目，并指定要咨询的 model。参见 为会话指定顾问。
  • 现在你可以控制 Claude Managed Agents 智能体的模型推理运行位置。在创建智能体时，在 model 对象内设置 inference_geo，或者为单个会话覆盖该设置。可用地理区域和定价请参见数据驻留。
  • Claude Managed Agents 会话现在可以从 GitHub 仓库加载技能。当会话挂载仓库时，其根目录 .claude/skills 中的任何技能都会在会话启动时被自动发现，并在该会话中供智能体使用。

2026 年 8 月 5 日

  • 推理钩子现已在 Claude Enterprise 组织中进入测试阶段。将 Claude 指向你组织的 AI 安全服务器，claude.ai、Cowork 和 Claude Code 中每个受管控的提示词都会先等待服务器给出允许或拒绝的判定，然后才继续推理。请求经过签名，失败处理可配置，每次拒绝都会记录在合规活动信息流中。请参见推理钩子。
  • 我们已停用 Claude Opus 4.1 模型（claude-opus-4-1-20250805）。现在通过 Claude API 向该模型发出的所有请求都将返回错误。我们建议升级到 Claude Opus 5。研究人员可以通过外部研究人员访问计划申请持续访问权限。

2026 年 8 月 3 日

  • 合规 API 现在会返回在 claude.ai 网页版或移动端启动的 Cowork 会话记录，该功能在 Claude Enterprise 组织中处于测试阶段。GET /v1/compliance/apps/sessions/remote 列出会话，GET /v1/compliance/apps/sessions/remote/{session_id}/messages 返回单个会话的记录，使用你现有的合规访问密钥以及 read:compliance_user_data 权限范围。请参见云端会话。

2026 年 8 月 1 日

2026 年 7 月 24 日

  • 我们推出了 Claude Opus 5（claude-opus-5），相比 Claude Opus 4.8 实现了跨越式提升。Claude Opus 5 支持100 万 token 上下文窗口（默认值和最大值均为 100 万），最大输出 token 数为 128k，并默认开启思考，价格为每百万 token 5 美元 / 25 美元，与 Claude Opus 4.8 相同。它已在 Claude API、Amazon Bedrock 中的 Claude、AWS 上的 Claude Platform、Google Cloud 上的 Claude 以及 Microsoft Foundry 中的 Claude 上提供。新功能、行为变更和迁移指南请参见Claude Opus 5 的新增内容，完整规格请参见模型概览。
  • 在 Claude Opus 5 上，仅允许在 effort 为 high 或更低时禁用思考：在 effort 为 xhigh 或 max 时使用 thinking: {"type": "disabled"} 会返回 400 错误，这是相较 Claude Opus 4.8 的一项破坏性变更。请参见Claude Opus 5 的新增内容。
  • Effort 是引导 Claude Opus 5 的主要控制手段：该模型支持完整的档位阶梯（low、medium、high、xhigh、max），其中 max 适用于对能力要求极高的工作。
  • 对话中途更改工具现已在 Claude Fable 5、Claude Mythos 5、Claude Opus 4.8 和 Claude Opus 5 上进入测试阶段：在对话轮次之间添加或移除工具，同时保留提示词缓存。请在请求中包含 mid-conversation-tool-changes-2026-07-01 beta 标头。
  • fallbacks 参数现在支持 "default" 模式，该模式会按拒绝类别应用 Anthropic 推荐的备用模型。服务端备用处于测试阶段，且 "default" 模式需要 server-side-fallback-2026-07-01 beta 标头。请参见拒绝与备用。
  • 我们已移除 Claude Opus 4.7 的快速模式。使用 speed: "fast" 向 claude-opus-4-7 发起的请求现在会返回错误；与 Claude Opus 4.6 不同，它们不会回退到标准速度。Claude Opus 4.7 本身仍以标准速度提供。要继续使用快速模式，请迁移到 Claude Opus 5 或 Claude Opus 4.8。详见快速模式。

2026年7月22日

  • 现在你可以在 Claude Managed Agents 代理的模型配置上设置 effort 级别。在创建代理时，将 effort 传入 model 对象中。各等级的作用请参见努力等级。
  • Claude Managed Agents 的 Webhook 现在覆盖环境和内存存储生命周期：四种 environment.* 事件类型和三种 memory_store.* 事件类型。你无需轮询即可对环境及内存存储的生命周期变化做出响应。请参见订阅 Webhook 中的 Environment events 和 Memory store events 标签页。
  • 创建 Claude Managed Agents 会话时，你现在可以用初始事件为其播种。在 POST /v1/sessions 上传入 initial_events，最多包含 50 个 user.message 和 user.define_outcome 事件。非空列表会在同一次调用中启动代理循环，因此你无需单独发送事件请求即可开始工作。
  • 在更新 Claude Managed Agents 代理时，version 字段现在是可选的。提供它以进行乐观并发控制（不匹配会返回 409 错误），或省略它以无条件应用更新。请参见更新语义。
  • Claude Managed Agents 会话线程事件流现在支持事件增量。GET /v1/sessions/{session_id}/threads/{thread_id}/stream 接受与会话级流相同的 event_deltas[] 查询参数，因此你可以在模型生成时预览子代理的文本。一个连接只会预览它正在读取的线程。请参见预览会话线程事件。

2026年7月17日

  • Claude Console 中旧版 Workbench（platform.claude.com/workbench）即将停用，访问将于 2026 年 8 月 17 日终止。更新后的 Workbench 不支持已保存的提示词、变量和评估。你可以从横幅中以及 Organizational Settings 下导出任何想要保留的数据。更多信息请参见 Claude 帮助中心中的如何使用 Workbench？。
  • 用于生成、改进和模板化提示词的实验性提示词工具 API（/v1/experimental/generate_prompt、/v1/experimental/improve_prompt 和 /v1/experimental/templatize_prompt）将随 Workbench 一起于 2026 年 8 月 17 日停用。移除后，对这些端点的请求将返回错误。

2026年7月15日

2026年7月14日

  • 你现在可以使用 Admin API 管理 Claude Enterprise（claude.ai）组织中的人员，该 API 面向所有 Claude Enterprise 组织提供 beta 版：列出成员并通过电子邮件地址查找他们、更改成员角色、移除成员、发送和撤回邀请、管理群组及其成员资格，以及读取自定义角色。群组和自定义角色请求需要 anthropic-beta: ce-user-management-2026-07-13 beta 标头；成员和邀请请求不需要 beta 标头。具有 read:org_audit 范围的 Admin API 密钥还可以调用所有用户管理 GET 端点。请参见用户管理。

2026年7月10日

  • Dreams（研究预览版）现已支持 Claude Fable 5 和 Claude Sonnet 5。请参阅支持的模型。
  • 我们扩展了 Access Transparency 关于 cmek_preserve 事件的文档，新增了一个筛选示例、一个示例事件负载，以及两个保留原因代码（policy_violation_investigation、csae_report）。文档现在还说明，无论保留是由人工审核员还是自动化安全流水线发起，都会写入保留事件。请参阅 CMEK 内容保留。

2026 年 7 月 8 日

  • 现在，在 Claude Console 中创建 API 密钥或 Admin API 密钥时，你可以设置过期时间。可选择预设值、自定义时长或永不过期。对于有效期至少为 7 天的密钥，Anthropic 会在过期前通过电子邮件通知创建者。现有密钥不受影响。Admin API 会在 expires_at 字段中报告每个密钥的过期时间。请参阅身份验证。

2026 年 7 月 2 日

  • 我们新增了 agent-memory-2026-07-22 beta 标头，它改变了列出记忆（GET /v1/memory_stores/{memory_store_id}/memories）的行为：结果按服务器定义的稳定顺序返回，且 order_by 和 order 参数会被忽略；depth 仅接受 0、1 或省略（其他值会返回 400 错误）；并且 path_prefix 必须以 / 结尾，并匹配整个路径段而非子字符串。未使用该标头时发出的分页游标在该标头下无效，因此采用该标头时请从第一页重新开始。在记忆存储端点上，agent-memory-2026-07-22 取代 managed-agents-2026-04-01；同时发送两者会返回 400 错误。2026 年 7 月 22 日，managed-agents-2026-04-01 标头将采用相同的列表行为。请参阅Beta 标头。
  • Python（0.116.0）、TypeScript（0.110.0）、Go（1.56.0）、Java（2.48.0）、Ruby（1.55.0）、PHP（0.36.0）、C#（12.35.0）和 CLI（1.16.0）SDK 现在在所有记忆存储调用中发送 agent-memory-2026-07-22，而非 managed-agents-2026-04-01。如果你的代码在记忆存储调用中显式传入 betas，请将那里的 managed-agents-2026-04-01 替换为 agent-memory-2026-07-22，而不是再添加第二个值。

2026 年 7 月 1 日

  • 我们已恢复对 Claude Fable 5 和 Claude Mythos 5 的访问。更多信息请参阅我们的声明。

2026 年 6 月 30 日

  • 我们推出了 Claude Sonnet 5（claude-sonnet-5），这是我们 Sonnet 模型系列的下一代产品，推出时定价为每百万 token 输入 $2 / 输出 $10（该价格于 2026 年 8 月 10 日成为标准价格）。Claude Sonnet 5 支持 100 万 token 上下文窗口、最大 128k 输出 token，以及除 Priority Tier 之外与 Claude Sonnet 4.6 相同的工具和平台功能集，Priority Tier 在 Claude Sonnet 5 上不可用。迁移时有三项行为变更：自适应思考现在默认开启；手动扩展思考（thinking: {type: "enabled", budget_tokens: N}）已被移除并返回 400 错误（它已在 Sonnet 4.6 上弃用）；将采样参数（temperature、top_p、top_k）设置为非默认值会返回 400 错误。Claude Sonnet 5 还使用了新的分词器，对于相同文本会产生大约多 30% 的 token。确切的增幅取决于内容和工作负载形态。详情和迁移指南请参阅Claude Sonnet 5 的新增功能。有关行为差异和模型特定的提示模式，请参阅为 Claude Sonnet 5 编写提示。
  • Claude Managed Agents 会话事件流现在支持事件增量。在 GET /v1/sessions/{session_id}/events/stream 上使用 event_deltas[] 查询参数即可启用。event_start 和 event_delta 事件会在完整的 agent.message 事件到达之前，预览代理消息生成过程中的文本。
  • 列出会话功能现已支持向后分页。GET /v1/sessions 会返回一个 prev_page 游标以及 next_page；将其作为 page 参数传入即可返回上一页。请参阅分页。
  • 创建 Claude Managed Agents 会话时，你现在可以为该会话覆盖智能体的配置。传入 agent 和 type: "agent_with_overrides"，即可为单个会话替换模型、系统提示、工具、MCP 服务器或技能。智能体本身保持不变。
  • Claude Managed Agents 保险库现在支持在环境变量凭据（环境变量选项卡）上设置 injection_location。它控制凭据的值在出站时是否被替换到智能体的出站请求头、请求体，或两者中。
  • Webhooks for Claude Managed Agents now cover the agent, deployment, and deployment run lifecycle. You can react to a newly published agent version, a paused deployment, or a failed scheduled run without polling. See the Agent events, Deployment events, and Deployment run events tabs in Subscribe to webhooks.

June 29, 2026

  • We've removed fast mode for Claude Opus 4.6. Requests to claude-opus-4-6 with speed: "fast" no longer run at fast speed or premium pricing: they run at standard speed, are billed at standard rates, and do not return an error. The response's usage.speed field reports the speed used. To continue using fast mode, migrate to Claude Opus 4.8. Read more in Fast mode.

June 26, 2026

  • We've raised rate limits across the Claude API. Claude Sonnet and Claude Haiku rate limits now match Claude Opus at every usage tier, and usage tiers have been consolidated into three: Start, Build, and Scale. Most organizations move to a higher tier, no organization receives lower limits than before, and no action is required. You can view your tier and current limits in the Claude Console.

June 25, 2026

  • We've deprecated fast mode for Claude Opus 4.7, with removal on July 24, 2026. After removal, requests to claude-opus-4-7 with speed: "fast" will return an error. Migrate to fast mode for Claude Opus 4.8. Read more in Fast mode.

June 22, 2026

  • MCP tunnels (research preview): the management API moved from /v1/organizations/tunnels on the Admin API to /v1/tunnels on the Claude API. The new surface uses the anthropic-beta: mcp-tunnels-2026-06-22 header and the workspace:manage_tunnels WIF scope. The previous surface remains available during a migration window. See the Tunnels API reference.

June 18, 2026

  • The Python, TypeScript, Go, Java, Ruby, PHP, and C# SDKs now include support for code_execution_20260120, the code execution tool version that adds REPL state persistence and is the minimum version for programmatic tool calling. To adopt it, set the tool's type to code_execution_20260120; no beta header is required. It's available on Claude Fable 5, Claude Mythos 5, Claude Opus 4.5 and newer, and Claude Sonnet 4.5 and newer; see the code execution tool's Compatibility section.

June 15, 2026

  • We've retired the Claude Sonnet 4 model (claude-sonnet-4-20250514) and the Claude Opus 4 model (claude-opus-4-20250514). All requests to these models on the Claude API will now return an error. We recommend upgrading to Claude Sonnet 4.6 and Claude Opus 4.8 respectively. Researchers can request ongoing access through the External Researcher Access Program.

June 11, 2026

  • The code execution tool now supports code_execution_20260521, which discloses the 90-second per-cell execution time limit in the tool description so Claude can budget long-running cells. No beta header is required.
  • The web search tool and web fetch tool now support web_search_20260318 and web_fetch_20260318, adding a response_inclusion parameter to drop consumed result blocks from the API response for agentic workflows. No beta header is required.

June 10, 2026

June 9, 2026

  • We've launched Claude Fable 5 (claude-fable-5), our most capable model open to all customers, alongside Claude Mythos 5 (claude-mythos-5) for Project Glasswing participants. Both models support a 1M token context window by default, 128k max output tokens, and always-on adaptive thinking. See Introducing Claude Fable 5 and Claude Mythos 5 for capabilities, API changes, and availability.
  • Claude Fable 5 and Claude Mythos 5 use the tokenizer introduced with Claude Opus 4.7. Compared to models before Claude Opus 4.7, the same text produces roughly 30% more tokens. The exact increase depends on the content and workload shape. Use the token counting API with model: "claude-fable-5" to measure your prompts under the new tokenizer.
  • Claude Fable 5 runs safety classifiers on requests and during response generation. When a classifier declines a request, the Messages API returns stop_reason: "refusal". You are not billed for a request refused before any output is generated. An opt-in fallbacks parameter (in beta on the Claude API and Claude Platform on AWS; not supported on the Message Batches API) re-runs refused requests on another model, billed at the fallback model's rates. See Handling stop reasons.
  • The stop_details.category field on refusal responses now includes "reasoning_extraction" on Claude Fable 5, returned when a request is blocked under Anthropic's Terms of Service restrictions on reverse engineering or duplicating model outputs. The existing "cyber" and "bio" categories are unchanged. No beta header is required.
  • On Claude Fable 5 and Claude Mythos 5, adaptive thinking is the only thinking mode: thinking: {"type": "disabled"} is not supported, and manual extended thinking budgets and assistant prefill are not supported (both return a 400 error). See Migrating from Claude Mythos Preview to Claude Mythos 5.
  • On Claude Fable 5 and Claude Mythos 5, thinking.display defaults to "omitted", the same as Claude Opus 4.8, Claude Opus 4.7, and Claude Mythos Preview; set display: "summarized" to receive readable thinking summaries. The raw chain of thought is never returned; pass thinking blocks back unchanged in multi-turn conversations on the same model. See Thinking output on Claude Fable 5 and Claude Mythos 5.
  • Claude Fable 5 requires 30-day data retention and is not available under zero data retention. See Model-specific data retention requirements.
  • Claude Managed Agents now supports scheduled deployments, letting you run sessions on a cron schedule without managing your own scheduler.
  • Claude Managed Agents vaults now support environment variable credentials, so you can securely inject secrets into the agent's sandbox for CLIs, SDKs, and other services that authenticate through environment variables.
  • The Compliance API Activity Feed (GET /v1/compliance/activities) is now available on Claude Platform on AWS. See IAM actions for Claude Platform on AWS for the ListComplianceActivities action that authorizes it.
  • The session.thread_* webhook events now include a session_thread_id field identifying the multiagent thread that triggered the event.
  • We've released a Swift package in beta that adds Claude as a server-side LanguageModel in Apple's Foundation Models framework. Call Claude through the same LanguageModelSession API as Apple's on-device model on iOS 27, macOS 27, visionOS 27, and watchOS 27 (beta).

June 5, 2026

  • We announced the deprecation of the Claude Opus 4.1 model (claude-opus-4-1-20250805), with retirement on the Claude API scheduled for August 5, 2026. We recommend migrating to Claude Opus 4.8. Read more in Model deprecations.

June 2, 2026

  • The advisor tool now supports a max_tokens parameter to cap the advisor model's output per call, reducing latency and output token cost for workloads that don't need full-length advisor responses. Set tools[].max_tokens on the advisor tool definition; see Capping advisor output.
  • On the Claude API, you are no longer billed for a request when it returns stop_reason: "refusal" without Claude having generated any output. See Streaming refusals for detecting and handling refusals.

May 29, 2026

May 28, 2026

  • We've launched Claude Opus 4.8 (), our most capable model. Claude Opus 4.8 supports a 1M token context window by default on the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry, 128k max output tokens, and the same set of tools and platform features as Claude Opus 4.7. See the migration guide for baseline settings, features, and migration guidance.
  • We've launched mid-conversation system messages. On Claude Opus 4.8, you can send role: "system" messages after a user turn (subject to placement rules) in the messages array, preserving prompt cache hits when instructions change during a long-running session. No beta header is required.
  • The stop_details field on refusal responses is now publicly documented; it returns a category (cyber, bio, or null) and a human-readable explanation, so your application can route different classes of refusal to the right next step. No beta header is required.
  • On Claude Opus 4.8, the effort parameter defaults to high across all surfaces, including Claude Code and the Messages API.
  • On Claude Opus 4.8, the minimum cacheable prompt length for prompt caching is 1,024 tokens, lower than on Claude Opus 4.7.
  • With adaptive thinking enabled, Claude Opus 4.8 triggers reasoning only when a turn needs it, reducing wasted thinking tokens compared to Claude Opus 4.7 at the same effort level.
  • Claude Opus 4.8 supports high-resolution image input (up to 2576 pixels on the long edge), same as Claude Opus 4.7.
  • Task budgets now support Claude Opus 4.8.
  • The advisor tool now supports Claude Opus 4.8.
  • Computer use now supports Claude Opus 4.8.
  • Fast mode for Claude Opus 4.8 is available as a research preview on the Claude API only.
  • Setting the sampling parameters temperature, top_p, or top_k to a non-default value returns a 400 error on Claude Opus 4.8, same as on Claude Opus 4.7. See the migration guide for details.
  • In Claude Code, we've expanded Auto mode to more users for long-running tasks. See the Claude Code documentation.
  • In Claude Code, Max plan users now default to fast mode on Claude Opus 4.8. See the Claude Code documentation.
  • In Claude Code, Workflows are available as a research preview, letting you define and run multistep agentic plans. See the Claude Code documentation.
  • We've deprecated fast mode for Claude Opus 4.6, with removal approximately 30 days after launch. Migrate to fast mode for Claude Opus 4.8 or Claude Opus 4.7. Read more in Fast mode.
  • For updates to claude.ai, Cowork, Claude for Microsoft 365, and other Claude apps in this release, see the release notes for Claude Apps.

May 27, 2026

  • The Messages API response now includes usage.output_tokens_details.thinking_tokens, reporting how many of the billed output tokens were extended thinking. When streaming, the breakdown appears only on the final message_delta event. No beta header is required.

May 19, 2026

  • MCP tunnels is now available as a research preview, so you can connect to MCP servers in your private network.
  • Self-hosted sandboxes are now available for Claude Managed Agents, as an alternative to running tool execution in Anthropic's infrastructure. See Self-hosted sandboxes.
  • With Claude Managed Agents, you can now update the agent's MCP server and tool configurations associated with an active session.
  • With Claude Managed Agents, large outputs from agent_toolset and MCP tools exceeding 100K characters (about 25K tokens) are now automatically spilled to a file in the sandbox. The model receives a truncated preview with the file path and can read the full content from there.

May 18, 2026

  • The web search tool now returns richer SEC filing data, making it easier to ground financial research agents, earnings analysis, and due-diligence workflows in primary sources with citations.

May 13, 2026

  • We've launched cache diagnostics in public beta. Pass diagnostics.previous_message_id on a Messages request and the API reports a cache_miss_reason explaining where the prompt cache prefix diverged from the previous turn. Include the cache-diagnosis-2026-04-07 beta header in your requests.

May 12, 2026

  • Fast mode (research preview) now supports Claude Opus 4.7. Set speed: "fast" with model: "claude-opus-4-7" and the fast-mode-2026-02-01 beta header for significantly faster output token generation at premium pricing. Pricing, rate limits, and access are the same as for Opus 4.6 fast mode; interested customers should join the waitlist.

May 11, 2026

  • We've launched Claude Platform on AWS, bringing the Claude API to Anthropic-managed infrastructure accessible through AWS, with AWS billing and IAM authentication. Access the full Messages API, Files API, Message Batches API, Claude Managed Agents, Agent Skills, code execution, and tool use through native AWS endpoints. Learn more in Claude Platform on AWS.

May 6, 2026

  • Multiagent orchestration and Outcomes are now in public beta under the standard managed-agents-2026-04-01 beta header.
  • Claude Managed Agents vault credential background refresh is now supported for mcp_oauth credentials. See Authenticate with vaults.
  • Webhooks for Claude Managed Agents are now supported. Webhook event types include session and vault lifecycle events. See Subscribe to webhooks.
  • Additional filtering and sorting options are now supported for Claude Managed Agents. Sessions can be filtered by status, and events can be filtered by type. Events can now be filtered by creation time.
  • Dreams for Claude Managed Agents are now available as a research preview. A dream reads an existing memory store alongside past session transcripts and produces a reorganized output memory store with duplicates merged, stale entries replaced, and new insights surfaced. Dream endpoints are gated by the dreaming-2026-04-21 beta header. Request access to try it.

May 4, 2026

  • We've launched Workload Identity Federation. Authenticate workloads to the Claude API with short-lived OIDC tokens from your own identity provider (AWS IAM, Google Cloud, GitHub Actions, Kubernetes, Microsoft Entra ID, Okta, SPIFFE, and more) instead of long-lived static API keys. Configure issuers and federation rules in the Claude Console, and the SDK handles token exchange and refresh automatically. See Authentication.

April 30, 2026

  • We've retired the 1M token context window beta (context-1m-2025-08-07) for Claude Sonnet 4.5 and Claude Sonnet 4. The beta header now has no effect on these models, and requests exceeding the standard 200k-token context window return an error. To use the 1M context window, migrate to Claude Sonnet 4.6 or Claude Opus 4.6, where it's included at standard pricing with no beta header required.

April 29, 2026

  • We've released the Claude API skill, an open-source Agent Skill that gives Claude up-to-date reference material for building on the Messages API and Claude Managed Agents across 8 languages. The skill is bundled with Claude Code and available in the Anthropic skills repository.

April 24, 2026

  • We've released the Rate Limits API, allowing administrators to programmatically query the rate limits configured for their organization and workspaces.

April 23, 2026

  • Memory for Claude Managed Agents is now in public beta under the standard managed-agents-2026-04-01 header. See Using agent memory for the full integration guide.

April 20, 2026

  • We've retired the Claude Haiku 3 model (claude-3-haiku-20240307). All requests to this model will now return an error. We recommend upgrading to Claude Haiku 4.5.

April 16, 2026

  • We've launched Claude Opus 4.7, our most capable model for complex reasoning and agentic coding, at the same $5 / $25 per MTok pricing as Opus 4.6. See What's new in Claude Opus 4.7 for capability improvements, new features, and the updated tokenizer. Opus 4.7 includes API breaking changes versus Opus 4.6; see the migration guide before upgrading.
  • Claude in Amazon Bedrock is now open to all Amazon Bedrock customers. Claude Opus 4.7 and Claude Haiku 4.5 are available self-serve from the Bedrock console through the Messages API endpoint at /anthropic/v1/messages, in 27 AWS regions with global and regional endpoints.
  • We've launched task budgets in beta on Claude Opus 4.7. Give Claude an advisory token budget for a full agentic loop (thinking, tool calls, tool results, and output) and the model sees a running countdown, using it to prioritize work and finish gracefully as the budget is consumed. Include the task-budgets-2026-03-13 beta header in your requests.
  • Claude Opus 4.7 supports high-resolution image input, raising the maximum image resolution from 1568 to 2576 pixels on the long edge for improved performance on computer use, screenshot understanding, and document analysis. High-resolution support is automatic and requires no beta header; images may use up to approximately 3x more image tokens than on prior models.
  • We've added the xhigh effort level on Claude Opus 4.7. xhigh sits between high and max and is tuned for long-running agentic and coding tasks (over 30 minutes) with token budgets in the millions. No beta header is required.

April 14, 2026

  • We announced the deprecation of the Claude Sonnet 4 model (claude-sonnet-4-20250514) and the Claude Opus 4 model (claude-opus-4-20250514), with retirement on the Claude API scheduled for June 15, 2026. We recommend migrating to Claude Sonnet 4.6 and Claude Opus 4.8 respectively. Read more in Model deprecations.

April 9, 2026

  • We've launched the advisor tool in public beta. Pair a faster executor model with a higher-intelligence advisor model that provides strategic guidance mid-generation, so long-horizon agentic workloads get close to advisor-solo quality while the bulk of token generation happens at executor-model rates. Include the beta header advisor-tool-2026-03-01 in your requests.

April 8, 2026

  • We've launched Claude Managed Agents in public beta, a fully managed agent harness for running Claude as an autonomous agent with secure sandboxing, built-in tools, and server-sent event streaming. Create agents, configure containers, and run sessions through the API. All endpoints require the managed-agents-2026-04-01 beta header. Learn more in Claude Managed Agents overview.
  • We've launched the ant CLI, a command-line client for the Claude API that enables faster interaction with the Claude API, native integration with Claude Code, and versioning of API resources in YAML files. Learn more in the CLI quickstart.

April 7, 2026

  • We announced Claude Mythos Preview is available as a gated research preview for defensive cybersecurity work as part of Project Glasswing. Access is invitation-only.
  • The Messages API is now available on Amazon Bedrock as a research preview. The new Claude in Amazon Bedrock endpoint at /anthropic/v1/messages uses the same request shape as the first-party Claude API and runs on AWS-managed infrastructure with zero operator access. Available in us-east-1; contact your Anthropic account executive to request access. Learn more in Claude in Amazon Bedrock.

March 30, 2026

  • We've raised the max_tokens cap to 300k on the Message Batches API for Claude Opus 4.6 and Sonnet 4.6. Include the output-300k-2026-03-24 beta header to generate longer single-turn outputs for long-form content, structured data, and large code generation tasks.
  • We're retiring the 1M token context window beta for Claude Sonnet 4.5 and Claude Sonnet 4 on April 30, 2026. After that date, the context-1m-2025-08-07 beta header will have no effect on these models, and requests that exceed the standard 200k-token context window will return an error. To continue using 1M context windows, migrate to Claude Sonnet 4.6 or Claude Opus 4.6, which support the full 1M token context window at standard pricing with no beta header required.

March 18, 2026

  • We've added model capability fields to the Models API. GET /v1/models and GET /v1/models/{model_id} now return max_input_tokens, max_tokens, and a capabilities object. Query the API to discover what each model supports.

March 16, 2026

  • We've launched the display field for extended thinking, letting you omit thinking content from responses for faster streaming. Set thinking.display: "omitted" to receive thinking blocks with an empty thinking field and the signature preserved for multi-turn continuity. Billing is unchanged. Learn more in Controlling thinking display.

March 13, 2026

  • The 1M token context window is out of beta for Claude Opus 4.6 and Sonnet 4.6, at standard pricing. Requests over 200k tokens work automatically for these models with no beta header required. The 1M token context window remains in beta for Claude Sonnet 4.5 and Sonnet 4.
  • We've removed the dedicated 1M rate limits for all supported models. Your standard account limits now apply across every context length.
  • We've raised the media limit from 100 to 600 images or PDF pages per request when using the 1M token context window.

February 19, 2026

  • We've launched automatic caching for the Messages API. Add a single cache_control field to your request body and the system automatically caches the last cacheable block, moving the cache point forward as conversations grow. No manual breakpoint management required. Works alongside existing block-level cache control for fine-grained optimization. Available on the Claude API and Microsoft Foundry (preview). Learn more in Prompt caching.
  • We've retired the Claude Sonnet 3.7 model (claude-3-7-sonnet-20250219) and the Claude Haiku 3.5 model (claude-3-5-haiku-20241022). All requests to Claude Sonnet 3.7 will now return an error. Requests to Claude Haiku 3.5 on the Claude API will now return an error; it remains available on Amazon Bedrock and Google Cloud. We recommend upgrading to Claude Sonnet 4.6 and Claude Haiku 4.5 respectively. Researchers can request ongoing access through the External Researcher Access Program.
  • We announced the deprecation of the Claude Haiku 3 model (claude-3-haiku-20240307), with retirement scheduled for April 20, 2026. We recommend migrating to Claude Haiku 4.5. Read more in Model deprecations.

February 17, 2026

February 7, 2026

  • We've launched fast mode in research preview for Opus 4.6, providing significantly faster output token generation through the speed parameter. Fast mode is up to 2.5x as fast at premium pricing. Interested customers should join the waitlist.

February 5, 2026

  • We've launched Claude Opus 4.6, our most intelligent model for complex agentic tasks and long-horizon work. Opus 4.6 recommends adaptive thinking (thinking: {type: "adaptive"}); manual thinking (type: "enabled" with budget_tokens) is deprecated. Opus 4.6 does not support prefilling assistant messages. Learn more in What's new in Claude 4.6.
  • The effort parameter no longer requires a beta header and now supports Claude Opus 4.6. Effort replaces budget_tokens for controlling thinking depth on new models.
  • We've launched the compaction API in beta, providing server-side context summarization for effectively infinite conversations. Available on Opus 4.6.
  • We've introduced data residency controls, allowing you to specify where model inference runs with the inference_geo parameter. US-only inference is available at 1.1x pricing for models released after February 1, 2026.
  • The 1M token context window is now available in beta for Claude Opus 4.6, in addition to Sonnet 4.5 and Sonnet 4. Long context pricing applies to requests exceeding 200k input tokens.
  • Fine-grained tool streaming no longer requires a beta header on any model or platform.

January 29, 2026

  • Structured outputs are out of beta on the Claude API for Claude Sonnet 4.5, Claude Opus 4.5, and Claude Haiku 4.5. This release includes expanded schema support, improved grammar compilation latency, and a simplified integration path with no beta header required. The output_format parameter has moved to output_config.format. Existing beta users can continue using the beta header during the transition period. Structured outputs remain in public beta on Amazon Bedrock and Microsoft Foundry.

January 12, 2026

  • console.anthropic.com now redirects to platform.claude.com. The Claude Console has moved to its new home as part of our Claude brand consolidation. Existing bookmarks and links will continue working through an automatic redirect. For more details, see the September 16, 2025 announcement.

January 5, 2026

  • We've retired the Claude Opus 3 model (claude-3-opus-20240229). All requests to this model will now return an error. We recommend upgrading to Claude Opus 4.5, which offers significantly improved intelligence at a third of the cost. Researchers can request ongoing access to Claude Opus 3 on the API through the External Researcher Access Program.

December 19, 2025

  • We announced the deprecation of the Claude Haiku 3.5 model. Read more in Model deprecations.

December 4, 2025

November 24, 2025

  • We've launched Claude Opus 4.5, our most intelligent model combining maximum capability with practical performance. Ideal for complex specialized tasks, professional software engineering, and advanced agents. Features step-change improvements in vision, coding, and computer use at a more accessible price point than previous Opus models. Learn more in Models overview.
  • We've launched programmatic tool calling in public beta, allowing Claude to call tools from within code execution to reduce latency and token usage in multi-tool workflows.
  • We've launched the tool search tool in public beta, enabling Claude to dynamically discover and load tools on-demand from large tool catalogs.
  • We've launched the effort parameter in public beta for Claude Opus 4.5, allowing you to control token usage by trading off between response thoroughness and efficiency.
  • We've added client-side compaction to our Python and TypeScript SDKs, automatically managing conversation context through summarization when using tool_runner.

November 21, 2025

  • Search result content blocks are now available on Amazon Bedrock with no beta header required. Learn more in Search results.

November 19, 2025

  • We've launched a new documentation platform at platform.claude.com/docs. Our documentation now lives side by side with the Claude Console, providing a unified developer experience. The previous docs site at docs.claude.com will redirect to the new location.

November 18, 2025

  • We've launched Claude in Microsoft Foundry, bringing Claude models to Azure customers with Azure billing and OAuth authentication. Access the full Messages API including extended thinking, prompt caching (5-minute and 1-hour), PDF support, Files API, Agent Skills, and tool use. Learn more in Claude in Microsoft Foundry.

November 14, 2025

  • We've launched structured outputs in public beta, providing guaranteed schema conformance for Claude's responses. Use JSON outputs for structured data responses or strict tool use for validated tool inputs. Available for Claude Sonnet 4.5 and Claude Opus 4.1. To enable, use the beta header structured-outputs-2025-11-13.

October 28, 2025

  • We announced the deprecation of the Claude Sonnet 3.7 model. Read more in Model deprecations.
  • We've retired the Claude Sonnet 3.5 models. All requests to these models will now return an error.
  • We've expanded context editing with thinking block clearing (clear_thinking_20251015), enabling automatic management of thinking blocks. Learn more in Context editing.

October 16, 2025

  • We've launched Agent Skills (skills-2025-10-02 beta), a new way to extend Claude's capabilities. Skills are organized folders of instructions, scripts, and resources that Claude loads dynamically to perform specialized tasks. The initial release includes:
    • Anthropic-managed Skills: Pre-built Skills for working with PowerPoint (.pptx), Excel (.xlsx), Word (.docx), and PDF files
    • Custom Skills: Upload your own Skills through the Skills API (/v1/skills endpoints) to package domain expertise and organizational workflows
    • Skills require the code execution tool to be enabled
    • Learn more in Agent Skills and API reference

October 15, 2025

  • We've launched Claude Haiku 4.5, our fastest and most intelligent Haiku model with near-frontier performance. Ideal for real-time applications, high-volume processing, and cost-sensitive deployments requiring strong reasoning. Learn more in Models overview.

September 29, 2025

  • We've launched Claude Sonnet 4.5, our best model for complex agents and coding, with the highest intelligence across most tasks. Learn more in the models overview.
  • We've introduced global endpoint pricing for Amazon Bedrock and Vertex AI. The Claude API (1P) pricing is unaffected.
  • We've introduced a new stop reason model_context_window_exceeded that allows you to request the maximum possible tokens without calculating input size. Learn more in Handling stop reasons.
  • We've launched the memory tool in beta, enabling Claude to store and consult information across conversations. Learn more in Memory tool.
  • We've launched context editing in beta, providing strategies to automatically manage conversation context. The initial release supports clearing older tool results and calls when approaching token limits. Learn more in Context editing.

September 17, 2025

  • We've launched tool helpers in beta for the Python and TypeScript SDKs, simplifying tool creation and execution with type-safe input validation and a tool runner for automated tool handling in conversations. For details, see the documentation for the Python SDK and the TypeScript SDK.

September 16, 2025

  • We've unified our developer offerings under the Claude brand. You should see updated naming and URLs across our platform and documentation, but our developer interfaces will remain the same. Here are some notable changes:

September 10, 2025

  • We've launched the web fetch tool in beta, allowing Claude to retrieve full content from specified web pages and PDF documents. Learn more in Web fetch tool.
  • We've launched the Claude Code Analytics API, enabling organizations to programmatically access daily aggregated usage metrics for Claude Code, including productivity metrics, tool usage statistics, and cost data.

September 8, 2025

  • We launched a beta version of the C# SDK.

September 5, 2025

  • We've launched rate limit charts in the Console Usage page, allowing you to monitor your API rate limit usage and caching rates over time.

September 3, 2025

  • We've launched support for citable documents in client-side tool results. Learn more in Handle tool calls.

September 2, 2025

  • We've launched v2 of the Code Execution Tool in public beta, replacing the original Python-only tool with Bash command execution and direct file manipulation capabilities, including writing code in other languages.

August 27, 2025

  • We launched a beta version of the PHP SDK.

August 26, 2025

August 19, 2025

  • Request IDs are now included directly in error response bodies alongside the existing request-id header. Learn more in Errors.

August 18, 2025

  • We've released the Usage & Cost API, allowing administrators to programmatically monitor their organization's usage and cost data.
  • We've added a new endpoint to the Admin API for retrieving organization information. For details, see the Organization Info Admin API reference.

August 13, 2025

  • We announced the deprecation of the Claude Sonnet 3.5 models (claude-3-5-sonnet-20240620 and claude-3-5-sonnet-20241022). These models will be retired on October 28, 2025. We recommend migrating to Claude Sonnet 4.5 (claude-sonnet-4-5-20250929) for improved performance and capabilities. Read more in Model deprecations.
  • The 1-hour cache duration for prompt caching no longer requires a beta header. Learn more in Prompt caching.

August 12, 2025

  • We've launched beta support for a 1M token context window in Claude Sonnet 4 on the Claude API and Amazon Bedrock.

August 11, 2025

  • Some customers might encounter 429 (rate_limit_error) errors following a sharp increase in API usage due to acceleration limits on the API. Previously, 529 (overloaded_error) errors would occur in similar scenarios.

August 8, 2025

  • Search result content blocks are out of beta on the Claude API and Vertex AI. This feature enables natural citations for RAG applications with proper source attribution. The beta header search-results-2025-06-09 is no longer required. Learn more in Search results.

August 5, 2025

  • We've launched Claude Opus 4.1, an incremental update to Claude Opus 4 with enhanced capabilities and performance improvements.* Learn more in Models overview.

*Opus 4.1 does not allow both temperature and top_p parameters to be specified. Please use only one.

July 28, 2025

  • We've released text_editor_20250728, an updated text editor tool that fixes some issues from the previous versions and adds an optional max_characters parameter that allows you to control the truncation length when viewing large files.

July 24, 2025

  • We've increased rate limits for Claude Opus 4 on the Claude API to give you more capacity to build and scale with Claude. For customers with usage tier 1-4 rate limits, these changes apply immediately to your account - no action needed.

July 21, 2025

  • We've retired the Claude 2.0, Claude 2.1, and Claude Sonnet 3 models. All requests to these models will now return an error. Read more in Model deprecations.

July 17, 2025

  • We've increased rate limits for Claude Sonnet 4 on the Claude API to give you more capacity to build and scale with Claude. For customers with usage tier 1-4 rate limits, these changes apply immediately to your account - no action needed.

July 3, 2025

  • We've launched search result content blocks in beta, enabling natural citations for RAG applications. Tools can now return search results with proper source attribution, and Claude will automatically cite these sources in its responses - matching the citation quality of web search. This eliminates the need for document workarounds in custom knowledge base applications. Learn more in Search results. To enable this feature, use the beta header search-results-2025-06-09.

June 30, 2025

  • We announced the deprecation of the Claude Opus 3 model. Read more in Model deprecations.

June 23, 2025

  • Console users with the Developer role can now access the Cost page. Previously, the Developer role allowed access to the Usage page, but not the Cost page.

June 11, 2025

  • We've launched fine-grained tool streaming in public beta, a feature that enables Claude to stream tool use parameters without buffering / JSON validation. To enable fine-grained tool streaming, use the beta header fine-grained-tool-streaming-2025-05-14.

May 22, 2025

  • We've launched Claude Opus 4 and Claude Sonnet 4, our latest models with extended thinking capabilities. Learn more in Models overview.
  • The default behavior of extended thinking in Claude 4 models returns a summary of Claude's full thinking process, with the full thinking encrypted and returned in the signature field of thinking block output.
  • We've launched interleaved thinking in public beta, a feature that enables Claude to think in between tool calls. To enable interleaved thinking, use the beta header interleaved-thinking-2025-05-14.
  • We've launched the Files API in public beta, enabling you to upload files and reference them in the Messages API and code execution tool.
  • We've launched the Code execution tool in public beta, a tool that enables Claude to execute Python code in a secure, sandboxed environment.
  • We've launched the MCP connector in public beta, a feature that allows you to connect to remote MCP servers directly from the Messages API.
  • To increase answer quality and decrease tool errors, we've changed the default value for the top_p nucleus sampling parameter in the Messages API from 0.999 to 0.99 for all models. To revert this change, set top_p to 0.999. Additionally, when extended thinking is enabled, you can now set top_p to values between 0.95 and 1.
  • Our Go SDK has moved from beta to its first stable release.
  • We've included minute and hour level granularity to the Usage page of Console alongside 429 error rates on the Usage page.

May 21, 2025

  • Our Ruby SDK has moved from beta to its first stable release.

May 7, 2025

  • We've launched a web search tool in the API, allowing Claude to access up-to-date information from the web. Learn more in Web search tool.

May 1, 2025

  • Cache control must now be specified directly in the parent content block of tool_result and document.source. For backwards compatibility, if cache control is detected on the last block in tool_result.content or document.source.content, it will be automatically applied to the parent block instead. Cache control on any other blocks within tool_result.content and document.source.content will result in a validation error.

April 9th, 2025

  • We launched a beta version of the Ruby SDK.

March 31st, 2025

  • Our Java SDK has moved from beta to its first stable release.
  • We've moved our Go SDK from alpha to beta.

February 27th, 2025

  • We've added URL source blocks for images and PDFs in the Messages API. You can now reference images and PDFs directly through a URL instead of having to base64-encode them. Learn more in Vision and PDF support.
  • We've added support for a none option to the tool_choice parameter in the Messages API that prevents Claude from calling any tools. Additionally, you're no longer required to provide any tools when including tool_use and tool_result blocks.
  • We've launched an OpenAI-compatible API endpoint, allowing you to test Claude models by changing just your API key, base URL, and model name in existing OpenAI integrations. This compatibility layer supports core chat completions functionality. Learn more in OpenAI SDK compatibility.

February 24th, 2025

  • We've launched Claude Sonnet 3.7, our most intelligent model yet. Claude Sonnet 3.7 can produce near-instant responses or show its extended thinking step-by-step. One model, two ways to think. Learn more about all Claude models in Models overview.
  • We've added vision support to Claude Haiku 3.5, enabling the model to analyze and understand images.
  • We've released a token-efficient tool use implementation, improving overall performance when using tools with Claude. Learn more in Tool use with Claude.
  • We've changed the default temperature in the Console for new prompts from 0 to 1 for consistency with the default temperature in the API. Existing saved prompts are unchanged.
  • We've released updated versions of our tools that decouple the text edit and bash tools from the computer use system prompt:
    • bash_20250124: Same functionality as previous version but is independent from computer use. Does not require a beta header.
    • text_editor_20250124: Same functionality as previous version but is independent from computer use. Does not require a beta header.
    • computer_20250124: Updated computer use tool with new command options including "hold_key", "left_mouse_down", "left_mouse_up", "scroll", "triple_click", and "wait". This tool requires the "computer-use-2025-01-24" anthropic-beta header. Learn more in Tool use with Claude.

February 10th, 2025

  • We've added the anthropic-organization-id response header to all API responses. This header provides the organization ID associated with the API key used in the request.

January 31st, 2025

  • We've moved our Java SDK from alpha to beta.

January 23rd, 2025

  • We've launched citations capability in the API, allowing Claude to provide source attribution for information. Learn more in Citations.
  • We've added support for plain text documents and custom content documents in the Messages API.

January 21st, 2025

  • We announced the deprecation of the Claude 2, Claude 2.1, and Claude Sonnet 3 models. Read more in Model deprecations.

January 15th, 2025

  • We've updated prompt caching to be easier to use. Now, when you set a cache breakpoint, we'll automatically read from your longest previously cached prefix.
  • You can now put words in Claude's mouth when using tools.

January 10th, 2025

December 19th, 2024

December 17th, 2024

The following features are now available in the Claude API without a beta header:

  • Models API: Query available models, validate model IDs, and resolve model aliases to their canonical model IDs.
  • Message Batches API: Process large batches of messages asynchronously at 50% of the standard API cost.
  • Token counting API: Calculate token counts for Messages before sending them to Claude.
  • Prompt Caching: Reduce costs by up to 90% and latency by up to 80% by caching and reusing prompt content.
  • PDF support: Process PDFs to analyze both text and visual content within documents.

We also released new official SDKs:

December 4th, 2024

November 21st, 2024

  • We've released the Admin API, allowing users to programmatically manage their organization's resources.

November 20th, 2024

  • We've updated our rate limits for the Messages API. We've replaced the tokens per minute rate limit with new input and output tokens per minute rate limits. Read more in Rate limits.
  • We've added support for tool use in the Workbench.

November 13th, 2024

  • We've added PDF support for all Claude Sonnet 3.5 models. Read more in PDF support.

November 6th, 2024

November 4th, 2024

November 1st, 2024

  • We've added PDF support for use with the new Claude Sonnet 3.5. Read more in PDF support.
  • We've also added token counting, which allows you to determine the total number of tokens in a Message prior to sending it to Claude. Read more in Token counting.

October 22nd, 2024

  • We've added Anthropic-defined computer use tools to our API for use with the new Claude Sonnet 3.5. Read more in Computer use tool.
  • Claude Sonnet 3.5, our most intelligent model yet, just got an upgrade and is now available on the Claude API. Read more in the Claude Sonnet documentation.

October 8th, 2024

  • The Message Batches API is now available in beta. Process large batches of queries asynchronously in the Claude API for 50% less cost. Read more in Batch processing.
  • We've loosened restrictions on the ordering of user/assistant turns in our Messages API. Consecutive user/assistant messages will be combined into a single message instead of erroring, and we no longer require the first input message to be a user message.
  • We've deprecated the Build and Scale plans in favor of a standard feature suite (formerly referred to as Build), along with additional features that are available through sales. Read more in our API pricing information.

October 3rd, 2024

  • We've added the ability to disable parallel tool use in the API. Set disable_parallel_tool_use: true in the tool_choice field to ensure that Claude uses at most one tool. Read more in Parallel tool use.

September 10th, 2024

  • We've added Workspaces to the Developer Console. Workspaces allow you to set custom spend or rate limits, group API keys, track usage by project, and control access with user roles. Read more in our blog post.

September 4th, 2024

August 22nd, 2024

  • We've added support for usage of the SDK in browsers by returning CORS headers in the API responses. Set dangerouslyAllowBrowser: true in the SDK instantiation to enable this feature.

August 19th, 2024

  • 8,192-token outputs on Claude Sonnet 3.5 are out of beta and no longer require the max-tokens-3-5-sonnet-2024-07-15 header.

August 14th, 2024

  • Prompt caching is now available as a beta feature in the Claude API. Cache and re-use prompts to reduce latency by up to 80% and costs by up to 90%.

July 15th, 2024

  • Generate outputs up to 8,192 tokens in length from Claude Sonnet 3.5 with the new anthropic-beta: max-tokens-3-5-sonnet-2024-07-15 header.

July 9th, 2024

  • Automatically generate test cases for your prompts using Claude in the Developer Console.
  • Compare the outputs from different prompts side by side in the new output comparison mode in the Developer Console.

June 27th, 2024

June 20th, 2024

  • Claude Sonnet 3.5, our most intelligent model yet, is now available across the Claude API, Amazon Bedrock, and Vertex AI.

May 30th, 2024

  • Tool use is out of beta across the Claude API, Amazon Bedrock, and Vertex AI, with no beta header required.

May 10th, 2024

  • Our prompt generator tool is now available in the Developer Console. Prompt Generator makes it easy to guide Claude to generate a high-quality prompts tailored to your specific tasks. Read more in our blog post.

来源：Claude Platform：开发者版本说明 · platform.claude.com

Anthropic / Claude模型发布产品更新#产品更新#模型发布#Anthropic