在 Claude 与 Claude Code 中充分发挥 Opus 5.5 的效能

Opus 5.5 与你现有的 Claude 使用方式配合良好，但也有几个行为不同：它能够更长时间地自主工作，会清楚说明自己做了什么，并且会在每次回复前进行思考。本指南介绍如何在 Claude 应用和 Claude Code 中使用 Opus 5.5，包括如何提示模型、引导一次长时间运行，以及检查结果。

先试试这些

你在首次使用 Opus 5.5 时可以尝试的三件事

1. 把整个任务交给它。说明“完成”是什么样子，以及你希望它何时停止并询问。然后让它自行工作。
2. 删除“仔细思考”之类的句子。Opus 5.5 已经会在每次回复前思考。
3. 一次长时间运行结束后，先阅读它需要你提供什么。

1. 如何提问

说明“完成”是什么样子，然后让它运行

该怎么做。 在一条消息中交代整个任务。说明完成标准，例如“测试全部通过”或“每个端点都已迁移”。然后让它自行处理。

为何这对 Opus 5.5 很重要。 Opus 5.5 在处理长时间、多部分工作时比 Opus 5 更能持续推进。与此前的 Opus 模型相比，它最大的提升体现在多步骤工作上，例如在大型代码库中持续完成一项变更，直到测试通过。早期测试者让它在几乎无需监督的情况下运行长时间编码任务数小时。有了明确的完成标准，它就知道何时任务结束。

如何做。 例如，在 Claude Code 中：

提示词复制

Migrate the payment endpoints from the old client to the new one. Done means: every endpoint uses the new client, the old client is deleted, and the test suite passes. Stop and ask me only if a test fails for a reason you can't explain.

图 A 一条消息包含：整个任务、完成标准，以及何时停止。

不要再要求它“努力思考”

该怎么做。 从提示词和已保存的指令中移除“仔细思考”“逐步思考”以及类似语句。

为何这对 Opus 5.5 很重要。 Opus 5.5 总会在回复前思考，并自行决定思考多少。你无需要求它思考。我们在聊天产品中的测试显示，移除“仔细思考”这句话后，回复开始得更快，且质量没有明显下降。

如何做。 直接删除这句话。若需要对简单问题快速作答，就明确说明：“直接回答。”若要改变 Claude Code 中的思考量，请调整 effort。

向正在进行的任务补充内容

该怎么做。 如果你在运行过程中想起某件事，可以在它工作时输入后续消息。

为何这对 Opus 5.5 很重要。 现在运行时间更长，重新开始的成本也更高。

如何做。 在 Claude Code 中，当 Claude 工作时输入消息并按 Enter，例如：“Also keep the old endpoint names as aliases.”

对于设计工作，说清你不想要哪些风格

该怎么做。 当你要求它制作网页、应用或其他产物时，列出你希望避免的设计习惯。

为何这对 Opus 5.5 很重要。 如果没有设计方向，Opus 5.5 会回退到少数几种默认风格。诸如“避免通用外观”这样的笼统指令，大多只是用一种默认风格替代另一种。列出具体模式效果好得多。

如何做。 点明这些模式：

提示词复制

Build a personal website with placeholder content. Don't use a cream or off-white background, italic accent words in headings, numbered "01 / 02 / 03" section labels, monospace labels, or pill-shaped buttons.

然后看看它改为选择了什么。如果你也不喜欢那些选择，就把它们补充到列表中，再次提出请求。

2. 在 CLAUDE CODE 中引导长时间运行

告诉它你希望在哪些情况下暂停

该怎么做。 在 CLAUDE.md 文件中加入一条简短规则，规定何时停止并询问，何时继续执行。

为何这对 Opus 5.5 很重要。 Opus 5.5 会在工作过程中持续向你通报进展。在长任务中，它有时会停下来汇报而不是继续：给出一份指出下一步但未采取该步骤的摘要、询问是否继续，或者列出实际上不阻碍工作的选项。它会遵循明确指出这些暂停情形的指令。因此，也请明确你希望它在哪些情况下暂停。

如何做。 将以下内容添加到 CLAUDE.md，并根据你的项目进行修改：

提示词复制

When a step doesn't need my input, keep going. Put status notes in the same message as your next action. Stop and ask only when you can't continue without me, or before anything destructive: deleting data, force-pushing, or changing anything outside this repository.

图 B CLAUDE.md 规则：何时继续执行，何时停止并询问。

如果一次运行以“Want me to continue?”结束，请回复“continue.”。如果这种情况经常发生，上述规则会有所帮助。

要求继续执行意味着更少暂停，因此对于任何有风险或难以撤销的操作，仍应保留你自己的检查环节。上面规则的最后一行就实现了这一点。对于破坏性命令，也请保持权限提示开启。

对于结对编程，你可能希望相反的行为：开始前先给出一行计划，结束后给出简短回顾。请在 CLAUDE.md 中改为这样说明。Opus 5.5 会遵循其中任何一种规则。

要求它将大型工作分配给子 Agent

该怎么做。 对于大型代码库中的审计、迁移或审查，要求 Opus 5.5 将工作拆分给子 Agent，并检查每项结果。

为何这对 Opus 5.5 很重要。 早期测试者曾让 Opus 5.5 在几乎无需监督的情况下，为长时间审计和迁移任务协调并行子 Agent。

如何做。

提示词复制

Audit every service in services/ for the retry bug in the linked issue. Give each service to its own subagent. When a subagent reports back, check its evidence before you accept it. Finish with one table: service, affected yes or no, and the evidence.

图 C 将工作分发给子 Agent，检查每个 Agent 的证据，然后用一张表收尾。

将任务列表保存在文件中

该怎么做。 对于需要一段时间的运行任务，要求 Opus 5.5 将任务列表保存在文件中，并在执行时持续更新。这样你可以阅读该文件而不是滚动历史记录，来查看运行进度。

为何这对 Opus 5.5 很重要。 现在运行时间更长。一次长时间运行会填满上下文窗口，Claude Code 随后会总结较早的对话轮次。文件中的列表不会受到影响，并可让你一眼看到哪些已完成、哪些尚未完成。

如何做。 “Keep a checklist in TASKS.md. Tick each item when it’s done, and add anything new you find.”

3. 检查结果

先阅读它需要你提供什么

该怎么做。 一次长时间运行结束后，先寻找 Claude 正等待你处理的事项，例如它留待你决定的问题，或希望你批准的变更。然后再阅读 Claude 的其余摘要。

为何这对 Opus 5.5 很重要。 Opus 5.5 比 Opus 5 更清楚地汇报其工作。它的更新和最终摘要会用直白语言说明：它做了什么、发现了什么，以及它需要你提供什么。

如何做。 若要更改摘要格式，请在 CLAUDE.md 中说明，例如：“End every run with three headings: Blocked on me, Changed, Found.”

要求它审查代码

该怎么做。 在人工审查之前，要求 Opus 5.5 审查 diff 或 pull request。

为何这对 Opus 5.5 很重要。 一名早期测试者表示，处于最低 effort 的 Opus 5.5 比高 effort 的 Opus 5 捕获到更多 bug，且误报更少。它还能用通俗语言解释自己的改动，因此它的 pull request 描述更容易审查。

如何做。 向 Claude 提交以下提示词：

提示词复制

Review the diff on this branch against main. List only problems you'd block the merge for. For each one, give the file and line, why it's wrong, and how to show it fails.

要求它标注无法确认的内容

该怎么做。 对于研究和分析任务，要求它说明哪些内容未能找到或无法核查。

为何这对 Opus 5.5 很重要。 “我找不到这个”值得阅读；主动要求提供这一信息，也能让它更容易被发现。

如何做。 在请求中加入“Mark anything you couldn’t confirm, and say where you looked”。这既适用于 Claude 研究报告，也适用于 Claude Code。

4. 在 CLAUDE 应用中

首先，确认模型选择器显示的是 Opus 5.5。

直接分享图表或截图本身

该怎么做。 附上图表、示意图、截图或幻灯片。不要重新输入其中的数字。

为何这对 Opus 5.5 很重要。 Opus 5.5 对图表、示意图和截图的读取比 Opus 5 更准确，而且不需要额外步骤。它也更擅长理解依赖图像中位置关系的含义：例如箭头连接了哪些方框、示意图两个版本之间改变了什么，或者日历截图中会议何时开始和结束。

如何做。 附上图片并提出具体问题：“Which of these services call the billing API directly?”

要求它检查长文档

该怎么做。 向它提供长篇计划、报告或演示文稿，并要求它找出错误。

为何这对 Opus 5.5 很重要。 Opus 5.5 比之前的 Opus 模型更关注细节。在我们的测试中，它在一个长篇规划对话中发现了日期对应错误的星期几，也发现了一张与演示文稿数字不一致的图表。

如何做。 提交以下提示词：“Check this deck for anything that contradicts itself: numbers, dates and names. Quote each problem and say where it is.”

要求提供完成后的文件

该怎么做。 当你需要电子表格或文档时，要求提供文件，而不是大纲。

为何这对 Opus 5.5 很重要。 Opus 5.5 制作的电子表格和文档，在分享之前需要的编辑工作比 Opus 5 制作的更少。

如何做。 “Make this a spreadsheet I can share: one row per vendor, with columns for cost, contract end date and owner.”

在项目中说明何时答案已经确定

该怎么做。 如果长对话中的后续问题感觉回复较慢，请添加一条指令，说明此前的答案已确定。

为何这对 Opus 5.5 很重要。 在长对话中，Opus 5.5 有时会在思考简短后续问题时重新审视先前的回答。这会拖慢回复速度。

如何做。 将以下内容添加到项目指令中：

提示词复制

Once you have answered something, treat that answer as done. Focus on what I'm asking now, and don't go back over an earlier answer unless I ask about it or point out a problem with it.

对于长篇分析项目，请不要添加这一规则，因为后续步骤可能会发现先前步骤中的错误。

5. 当消息被标记时

Opus 5.5 是首个以 Fable 级生物与网络安全防护上线的 Opus 模型。在 Claude 应用和 Claude Code 中，大多数被标记的消息会切换到旧模型，工作将在该模型上继续进行。在源代码中查找安全漏洞是被允许的，日常健康和教育问题也仍应可正常使用。这些防护措施有时会标记正当工作，我们正在调整它们以减少错误标记。如果发生切换，以下是你将看到的内容以及应采取的操作。

在 Claude 应用中

你会看到什么。 一则以“Switched to”开头、并显示某个旧模型名称的通知。Claude 会在该模型上回答，对话也会保持在该模型上。

该怎么做。

• 若要切回 Opus 5.5，请在模型选择器中选择它。如果较早的消息仍在对话中，它可能会再次被标记。新建对话可以避免这一情况。
• 若希望先收到询问，请前往 Settings，然后进入 Capabilities，关闭“Switch models when a message is flagged”。你会看到一张“paused”卡片，其中列出可选操作。

该检查覆盖对话中的所有内容，包括文件和搜索结果。因此，标记可能来自较早内容，而不只是你的最后一条消息。

在 Claude Code 中

你会看到什么。 一则显示旧模型名称的通知。会话将在该模型上继续。

该怎么做。

• 运行 /model 切回。
• 连按两次 Esc，编辑最后一条消息后重试。
• 若希望先收到询问，请运行 /config，然后修改“Switch models when a message is flagged”。
• 如果标记有误，请运行 /feedback。

不要要求它在回复中展示其推理过程

该怎么做。 从提示词和指令中移除要求它在回复中复现内部推理过程的内容。

为何这对 Opus 5.5 很重要。 要求在回复中复现内部推理过程，可能会被拒绝。这是标记类别之一。

如何做。 改为要求 Claude 提供你需要的内容，例如：“Explain why you chose this approach in three sentences.”

6. 速度

当你在等待每次回复时，开启快速模式

该怎么做。 在 Claude Code 中，对于你会在发送下一条消息前阅读每次回复的来回协作工作，请使用快速模式。

为何这对 Opus 5.5 很重要。 在发布时，Opus 5.5 已提供快速模式，作为

Opus 5.5 works well with the way you already use Claude. A few things behave differently, though: it works for longer on its own, it tells you plainly what it did, and it thinks before every reply. This guide covers how to work with Opus 5.5 in Claude apps and Claude Code, including how to prompt the model, steer a long run, and check your results.

TRY THIS FIRST

Three things to try in your first session with Opus 5.5

1. Hand over the whole task. Say what “done” looks like and when you want it to stop and ask. Then let it work.
2. Delete “think carefully” lines. Opus 5.5 already thinks before every reply.
3. When a long run ends, read what it needs from you first.

1. HOW TO ASK

Say what “done” looks like, then let it run

What to do. Give the whole task in one message. Name the finish line, like “the tests pass” or “every endpoint is migrated.” Then let it cook.

Why it matters on Opus 5.5. Opus 5.5 keeps going on long, multi-part work better than Opus 5 did. Compared to prior Opus models, its biggest gains are on multi-step work, like carrying a change through a large repository until the tests pass. Early testers had it run long coding tasks for hours with little oversight. With a clear finish line, it knows when it’s done.

How. In Claude Code, for example:

PROMPTCopy

Migrate the payment endpoints from the old client to the new one.
Done means: every endpoint uses the new client, the old client is deleted, and the test suite passes.
Stop and ask me only if a test fails for a reason you can't explain.

FIG A One message: the whole task, the finish line, and when to stop.

Stop telling it to “think hard”

What to do. Remove “think carefully,” “think step by step,” and similar lines from your prompts and your saved instructions.

Why it matters on Opus 5.5. Opus 5.5 always thinks before it replies, and it decides how much. You don’t need to ask it to think. In our testing in a chat product, removing a “think carefully” line made replies start sooner, with no clear drop in quality.

How. Delete the line. For a quick answer to a simple question, say so: “Answer directly.” To change how much it thinks in Claude Code, change effort.

Add to a running task

What to do. If you remember something mid-run, you can type a follow-up while it works.

Why it matters on Opus 5.5. Runs are longer now, so a restart costs more.

How to do it. In Claude Code, type the message and press Enter while Claude works, for example, “Also keep the old endpoint names as aliases.”

For design work, name the styles you don’t want

What to do. When you ask for a page, an app, or an artifact, list the design habits you want left out.

Why it matters on Opus 5.5. With no design direction, Opus 5.5 falls back on a few default styles. A general instruction like “avoid a generic look” mostly swaps one default for another. A list of specific patterns works much better.

How. Name the patterns:

PROMPTCopy

Build a personal website with placeholder content.
Don't use a cream or off-white background, italic accent words in headings, numbered "01 / 02 / 03" section labels, monospace labels, or pill-shaped buttons.

Then look at what it chose instead. If you don’t like that either, add it to the list and ask again.

2. STEERING A LONG RUN IN CLAUDE CODE

Tell it which stops you want

What to do. Put a short rule in your CLAUDE.md file about when to stop and ask, and when to keep going.

Why it matters on Opus 5.5. Opus 5.5 keeps you posted as it works. On a long task, it sometimes stops to report instead of going on: a summary that names the next step without taking it, an offer to continue, or a list of choices that don’t block the work. It follows instructions that name these stops. Name the stops you want, too.

How. Add this to CLAUDE.md, and edit it to fit your project:

PROMPTCopy

When a step doesn't need my input, keep going. Put status notes in the same message as your next action.
Stop and ask only when you can't continue without me, or before anything destructive: deleting data, force-pushing, or changing anything outside this repository.

FIG B The CLAUDE.md rule: when to keep going, and when to stop and ask.

If a run stops with “Want me to continue?” reply “continue.” If that happens often, the rule above will help.

A rule to keep going means fewer stops, so keep your own check before anything risky or hard to undo. The last line of the rule above does that. Keep permission prompts on for destructive commands too.

For pair programming, you may want the opposite: a one-line plan before it starts and a short recap at the end. Say that in your CLAUDE.md instead. Opus 5.5 follows either one.

Ask it to split big work across subagents

What to do. For an audit, a migration, or a review across a large codebase, ask Opus 5.5 to split the work across subagents and check each result.

Why it matters on Opus 5.5. Early testers had Opus 5.5 coordinate parallel subagents on long audits and migrations, with little oversight.

How.

PROMPTCopy

Audit every service in services/ for the retry bug in the linked issue.
Give each service to its own subagent. When a subagent reports back, check its evidence before you accept it.
Finish with one table: service, affected yes or no, and the evidence.

FIG C Fan out to subagents, check each one’s evidence, then finish with one table.

Keep the task list in a file

What to do. For a run that will take a while, ask Opus 5.5 to keep its task list in a file and update it as it goes. Then read the file, not the scrollback, to see where the run is.

Why it matters on Opus 5.5. Runs are longer now. A long run fills the context window, and Claude Code then summarizes older turns. A list in a file survives that, and it shows you at a glance what’s done and what’s left.

How. “Keep a checklist in TASKS.md. Tick each item when it’s done, and add anything new you find.”

3. CHECKING THE RESULT

Read what it needs from you first

What to do. When a long run ends, look first for anything Claude is waiting on you for, like a decision it left open or a change it wants you to approve. Then read the rest of Claude’s summary.

Why it matters on Opus 5.5. Opus 5.5 reports on its work more clearly than Opus 5. Its updates and its final summary say what it did, what it found, and what it needs from you, in plain language.

How. To change the summary’s format, say so in CLAUDE.md, for example, “End every run with three headings: Blocked on me, Changed, Found.”

Ask it to review the code

What to do. Ask Opus 5.5 to review a diff or a pull request before a person does.

Why it matters on Opus 5.5. One early tester said Opus 5.5 at its lowest effort caught more bugs than Opus 5 at high effort, with fewer false alarms. It also explains its changes in plain language, so its pull request descriptions are easier to review.

How. Feed this prompt to Claude:

PROMPTCopy

Review the diff on this branch against main.
List only problems you'd block the merge for. For each one, give the file and line, why it's wrong, and how to show it fails.

Ask it to mark what it couldn’t confirm

What to do. For research and analysis, ask it to say what it couldn’t find or couldn’t check.

Why it matters on Opus 5.5. “I couldn’t find this” is worth reading, and asking for it makes it easy to find.

How. Add “Mark anything you couldn’t confirm, and say where you looked” to the request. This works in a Claude research report and in Claude Code.

4. IN CLAUDE APPS

First, check that the model picker says Opus 5.5.

Share the chart or screenshot itself

What to do. Attach the chart, diagram, screenshot, or slide. Don’t retype the numbers.

Why it matters on Opus 5.5. Opus 5.5 reads charts, diagrams, and screenshots more accurately than Opus 5, and it needs no extra steps to do it. It’s also better at meaning that depends on where things are in the image: which boxes an arrow connects, what changed between two versions of a diagram, or when a meeting starts and ends in a calendar screenshot.

How. Attach the image and ask a specific question: “Which of these services call the billing API directly?”

Ask it to check a long document

What to do. Give it a long plan, report, or deck, and ask it to find mistakes.

Why it matters on Opus 5.5. Opus 5.5 pays more attention to detail than prior Opus models. In our testing, it caught a date that fell on the wrong weekday in a long planning thread, and a chart that didn’t match the numbers in a deck.

How. Submit the prompt: “Check this deck for anything that contradicts itself: numbers, dates and names. Quote each problem and say where it is.”

Ask for the finished file

What to do. When you want a spreadsheet or a document, ask for the file, not an outline.

Why it matters on Opus 5.5. The spreadsheets and documents Opus 5.5 makes need less editing than Opus 5’s before you share them.

How. “Make this a spreadsheet I can share: one row per vendor, with columns for cost, contract end date and owner.”

In a project, say when answers are settled

What to do. If follow-up questions in a long chat feel slow, add an instruction that earlier answers are settled.

Why it matters on Opus 5.5. In a long chat, Opus 5.5 sometimes goes back over an earlier answer while it thinks about a short follow-up. That slows the reply.

How. Add this to the project’s instructions:

PROMPTCopy

Once you have answered something, treat that answer as done. Focus on what I'm asking now, and don't go back over an earlier answer unless I ask about it or point out a problem with it.

Leave it out of projects for long analysis, where a later step can show a mistake in an earlier one.

5. WHEN A MESSAGE IS FLAGGED

Opus 5.5 is the first Opus model to launch with Fable-level bio and cyber safeguards. In Claude apps and Claude Code, most flagged messages move to an older model, and your work goes on there. Finding security vulnerabilities in source code is allowed, and everyday health and educational questions should still work. These safeguards can sometimes flag legitimate work, and we’re tuning them to cut down on incorrect flags. If you’re switched, here’s what you’ll see and what to do.

In Claude apps

What you see. A notice that starts with “Switched to” and the name of an older model. Claude answers on that model, and the chat stays on it.

What to do.

• To go back to Opus 5.5, choose it in the model picker. If the earlier message is still in the chat, it may be flagged again. Starting a new chat avoids that.
• To be asked first, go to Settings, then Capabilities, and turn off “Switch models when a message is flagged.” You’ll see a “paused” card with your options.

The check covers everything in the conversation, including files and search results. So a flag can come from earlier content, not only your last message.

In Claude Code

What you see. A notice that names the older model. The session continues on that model.

What to do.

• Run /model to switch back.
• Press Esc twice to edit your last message and try again.
• To be asked first, run /config and change “Switch models when a message is flagged.”
• Run /feedback if the flag was wrong.

Don’t ask it to show its reasoning in the reply

What to do. Remove requests to reproduce its internal reasoning in the reply from your prompts and instructions.

Why it matters on Opus 5.5. A request to reproduce its internal reasoning in the reply can be declined. It’s one of the flag categories.

How. Ask Claude for what you need instead, for example, “Explain why you chose this approach in three sentences.”

6. SPEED

Turn on fast mode when you’re waiting on each reply

What to do. In Claude Code, use fast mode for back-and-forth work, where you read each reply before you send the next message.

Why it matters on Opus 5.5. Fast mode is available for Opus 5.5 at launch as a research preview. You get the same model, and the text arrives sooner. It needs extra usage turned on, and it costs more per token than standard mode.

How. Type /fast into Claude.

YOUR OPUS 5.5 CHECKLIST

Run through this before your next long task.

FIG D The checklist at a glance.

Asking

• The task says what “done” looks like
• No “think hard” lines in prompts or saved instructions
• Design requests list the styles to leave out
• Charts and screenshots are attached, not retyped

Long runs in Claude Code

• CLAUDE.md says when to stop and when to keep going, and to stop before anything destructive
• Permission prompts are still on for destructive commands
• Large audits and migrations are split across subagents
• The task list is kept in a file

Checking

• The “needs from you” part of the report is read first
• A review pass runs before a person reviews
• Research answers mark what couldn’t be confirmed

Flags

• You know how to switch back: the model picker, or /model
• “Switch models when a message is flagged” is set the way you want

Start building with Opus 5.5!

With thanks to Molly Vorwerck for reviewing.