Modern Creator
Why Modern Creator?
Brock Mesarich | AI for Non Techies · YouTube

Anthropic's New Prompting Guide for Claude Opus 5.5

A creator speedruns Anthropic's own Opus 5.5 documentation into eight concrete changes, starting with a default effort setting that quietly got lower.

Posted
yesterday
Duration
Format
Listicle
educational
Views
7.7K
127 likes
Part of the collectionThe Claude Opus 5 PlaybookEvery Opus 5 breakdown, synthesized into one page.
Read the playbook
Big Idea

The argument in one line.

Anthropic's official Opus 5.5 documentation shows the model now defaults to a lower thinking effort than Opus 5, and most of the guide's other seven tips exist to help you stop overriding that default out of old habit.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • You already use Claude Opus for real work and want to update your prompting habits for the version change.
  • You maintain a long system prompt or reusable templates that were tuned for Opus 5's default thinking behavior.
  • You build agents or automations on Claude and need clearer rules for when a task actually counts as finished.
SKIP IF…
  • You've never used Claude Opus 5 or 5.5 and have no existing prompts to update.
  • You want a general prompting 101 video rather than a version-specific changelog.
TL;DR

The full version, fast.

Anthropic published an official prompting guide for Claude Opus 5.5, and this breakdown distills it into eight concrete changes. The headline shift: Opus 5.5 now defaults to medium thinking effort instead of Opus 5's high, so prompts built for the older model may be spending tokens they no longer need. The rest of the guide covers dropping old 'think carefully' instructions, making Claude check every relevant source before acting, separating pasted content from instructions, and defining what 'done' actually means for an unattended agent. The throughline: default settings changed, and habits tuned for Opus 5 need a deliberate reset, not a blind carryover.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →
Chapters

Where the time goes.

00:00 – 00:21

01 · Anthropic's Opus 5.5 prompting guide

A cold open frames the video: Anthropic just published an official prompting guide for Opus 5.5, and the host is going to speedrun the 8 key takeaways.

00:21 – 01:47

02 · Tip 1: The new default effort level

Opus 5 defaulted to high thinking effort; Opus 5.5 now defaults to medium. Anthropic recommends starting at medium and only raising effort once you've measured it helps your specific task, since the same effort label doesn't cost the same thinking across model versions.

01:47 – 03:13

03 · Tip 2: Drop the "think carefully" instructions

Instructions like 'think carefully before answering' may be redundant on Opus 5.5 because effort level already controls how much the model thinks. A vague thinking instruction is less useful than naming the exact comparison and tradeoff you want.

03:13 – 04:37

04 · Tip 3: Make Claude check every source

Opus 5.5 gets to work quickly, which can mean it acts before checking other relevant sources, like an email that moved a deadline. Anthropic recommends explicitly telling Claude to look through related documents before acting rather than assuming the first file is complete.

04:37 – 05:14

05 · Tip 4: Separate your request from pasted content

Pasting an email or webpage into a prompt mixes your instruction with someone else's content, which might contain its own embedded instructions. The fix is to state the job first, then paste the source material separately.

05:14 – 05:57

06 · Tip 5: Why Claude can seem silent

A missing progress update usually means the app isn't displaying messages Claude already sent, not that Claude is silent. The better fix is telling Claude exactly when you want an update.

05:57 – 06:56

07 · Tip 6: A finished reply isn't a finished task

A reply that sounds finished doesn't mean the task is done; a background step can still be running. For unattended agents, defining what 'done' includes lets Claude flag exactly what's still missing.

06:56 – 07:41

08 · Tip 7: Give Claude a clear design direction

Opus 5.5 still falls back on generic default website styles unless given concrete visual direction. Naming a specific background, headline style, button shape, and spacing beats a vague note like 'make it less generic.'

07:41 – 08:15

09 · Tip 8: Help Claude read small details

Opus 5.5 reads dense visuals and small labels more accurately than Opus 5 and can crop and inspect an image itself when tools are available. Pointing it to the exact detail that matters beats guessing.

08:15 – 08:40

10 · Final thoughts

A closing recap: pick one tip to try rather than stacking every recommendation into one prompt, since not all of them apply to every workflow.

Atomic Insights

Lines worth screenshotting.

  • Opus 5 defaulted to high thinking effort; Opus 5.5 now defaults to medium, so a system prompt built for the older model may be spending tokens it no longer needs to.
  • The same effort label doesn't cost the same amount of thinking across model versions, so carrying over an old effort setting can silently produce longer, pricier replies.
  • Anthropic's own guidance is to start at medium effort and only raise it once you've measured a real benefit on your own task, not by default.
  • Instructions like 'think carefully before answering' can be redundant on Opus 5.5, because the effort setting already controls how much the model thinks.
  • Removing 'think carefully' style instructions in Anthropic's own tests made replies start sooner with no measurable drop in quality.
  • Opus 5.5 can silently revisit an earlier answer while handling a new message, even a simple follow-up, unless told to treat prior answers as settled.
  • A vague prompt like 'think deeply and carefully' gives the model no real target; naming the exact comparison and tradeoff you want produces a more useful answer.
  • Opus 5.5 tends to act fast, which means it can miss context that lives outside the one file you handed it, like an email that moved a deadline.
  • Anthropic recommends explicitly telling Claude to check other likely sources before acting, not assume the first document it's handed is complete.
  • When you paste an email or webpage into a prompt, separate your instruction from the pasted content so Claude doesn't treat someone else's text as a command.
  • A missing status update usually means the app isn't displaying Claude's progress messages, not that Claude isn't producing them.
  • A finished-sounding reply from Opus 5.5 doesn't guarantee the task is done; a background step can still be running.
  • For unattended agents, defining what 'done' means, a report, a source list, a summary, lets Claude flag exactly what's missing instead of just being told to keep going.
  • Opus 5.5 still defaults to generic web design choices unless given concrete direction, so 'make it less generic' just swaps one default style for another.
  • Opus 5.5 reads dense visuals and small labels more accurately than Opus 5, and can crop and inspect an image itself if given the tools to do it.
Takeaway

Eight defaults that changed with Opus 5.5

WHAT TO LEARN

Anthropic's own Opus 5.5 guide is really a list of assumptions from the Opus 5 era that stopped being true, and the fix in each case is being more explicit, not working harder.

02Tip 1: The new default effort level
  • Opus 5.5's default effort dropped from high to medium, so a system prompt copied from Opus 5 may now be paying for more thinking than it needs.
  • Test effort levels against your own task before assuming you need the highest setting; a longer answer doesn't automatically mean a better one.
  • Changing the effort value between requests can invalidate the prompt cache, so switch it deliberately rather than per turn.
03Tip 2: Drop the "think carefully" instructions
  • Explicit instructions like 'think carefully before answering' can be redundant on Opus 5.5 since the effort setting already governs how much it thinks.
  • A vague instruction to think deeply gives the model no concrete target; naming the exact comparison and the tradeoff you want produces a sharper answer.
  • Opus 5.5 can revisit an earlier answer on a later turn even for a simple follow-up; tell it to treat prior answers as settled unless you're actually questioning them.
04Tip 3: Make Claude check every source
  • Opus 5.5 gets to work quickly, which means it can act before checking documents the request didn't explicitly name, like a related email or another spreadsheet tab.
  • Explicitly ask Claude to look through relevant sources before acting instead of assuming the first document it's handed contains everything.
  • This costs slightly more in tool calls and tokens, but it catches details like a deadline that moved somewhere the model wasn't told to look.
05Tip 4: Separate your request from pasted content
  • Pasted material can contain its own instructions that aren't actually yours; state your job first, then paste the source separately.
  • Making the boundary between your request and someone else's content explicit prevents Claude from treating pasted text as a command.
06Tip 5: Why Claude can seem silent
  • An app that doesn't display Claude's progress updates will look silent even when Claude is producing them; asking it to talk more won't fix a display problem.
  • Tell Claude when you actually want an update, like after files are reviewed or before a draft is sent, instead of leaving the timing open.
  • For automated setups, cap the number of reminders so updates give visibility without turning every step into an approval request.
07Tip 6: A finished reply isn't a finished task
  • A reply that sounds complete can still leave a background step running; check for actual completion instead of trusting the tone of the reply.
  • For unattended agents, define what 'done' means explicitly, a report, a source list, a summary, so Claude can flag exactly what's missing.
  • Reporting the next step is not the same as taking it; a checklist alone doesn't guarantee the work happened.
08Tip 7: Give Claude a clear design direction
  • Opus 5.5 still falls back on generic default styling; a vague note like 'make it less generic' just swaps one default for another.
  • Name a specific background, headline style, button shape, and spacing, or hand over a reference image, to get a result that isn't just the next default.
09Tip 8: Help Claude read small details
  • Opus 5.5 reads dense charts and small labels more accurately than Opus 5, and can crop and inspect an image itself when tools are available.
  • Point Claude to the exact label or region that matters rather than handing over the whole image and hoping it notices.
  • Ask Claude to say when a detail is unreadable instead of guessing; that's more useful than a confident wrong answer.
Glossary

Terms worth knowing.

Effort setting
A Claude Opus 5.5 parameter (low, medium, high) that controls how much internal reasoning the model does before answering. Higher effort means more thinking tokens and higher cost.
Thinking tokens
The tokens a model spends on internal reasoning before producing its visible answer. They count toward usage even though the user doesn't see that reasoning directly.
Prompt cache
A mechanism that reuses previously processed prompt content so repeated context doesn't get reprocessed at full price. Changing the top-level effort value between requests can invalidate it.
max_tokens
A request parameter that caps the total tokens a model can generate, including thinking tokens. Setting it too low can cut off a reply mid-thought.
Resources

Things they pointed at.

00:05linkAnthropic's official Opus 5.5 prompting guide
Quotables

Lines you could clip.

00:40
“The same effort level doesn't necessarily mean the same amount of thinking across these models.”
counterintuitive, contradicts the assumption that settings carry over between model versions→ TikTok hook↗ Tweet quote
01:27
“A longer answer alone doesn't necessarily tell you whether the effort level was worthwhile.”
short, standalone, reframes a common assumption→ IG reel cold open↗ Tweet quote
02:45
“When you write out a vague prompt, something like thinking deeply and carefully, it doesn't necessarily tell Claude what makes that answer useful to you.”
names a specific, common prompting mistake→ newsletter pull-quote↗ Tweet quote
04:13
“If Claude only reads the brief, it can produce a polished update with the wrong date, obviously.”
concrete failure example, easy to visualize→ TikTok hook↗ Tweet quote
05:35
“Asking Claude to talk more will not fix updates the app is hiding.”
tight, quotable, corrects a common misdiagnosis→ IG reel cold open↗ Tweet quote
07:18
“Anthropic says less generic often swaps out one default for another.”
punchy one-liner about vague design prompts→ newsletter pull-quote↗ Tweet quote
07:47
“Opus 5.5 reads visual material more accurately than Opus 5 without extra tools.”
direct capability claim, sourced from Anthropic→ TikTok hook↗ Tweet quote
The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

Anthropic just dropped this guide on how to use Opus 5 .5, and if you want to get the most out of it, I'm going to speedrun 8 tips covering the key takeaways. What's different, what to change, and practical examples you can use. And a few of these I had absolutely no idea about.
This should not only help you get better results with this model, but also save you money on token costs. So without further ado, let's dive right in. The first concrete change is the default effort setting.
Opus 5 defaulted to the high effort level, while Opus 5 .5 now defaults to medium basically effort controls how much work the model puts into thinking before it responds an entropic recommends starting at medium and testing it on your own tasks and don't assume that you need to carry over the higher setting that you used with the previous models the same effort level doesn't necessarily mean the same amount of thinking across these models if you keep your old configuration you could get longer turns and more output tokens entropic recommends lowering the effort level first when you want less things and then only using the highest settings where you've actually measured a benefit.
This way, it's going to save you a lot on token usage. Otherwise, you're just going to use a higher model, even though it might not be necessary with Opus 5 .5. And then for developers, the guide also covers leaving enough room for thinking tokens and changing effort without invalidating the prompt cache.
But if you're not a developer, this probably sounds like gibberish, so it probably doesn't apply to you. A useful way to choose your setting is to compare the same task at different effort levels. For example, you can give clod two proposals in the same list of requirements then compare both medium and the higher effort and then you could ask yourself did one miss an important requirement was the recommendation more useful how long did it take a longer answer alone doesn't necessarily tell you whether the effort level was worthwhile and this is going to be custom to you and your individual use cases all right moving on to tip number two and this is thinking instructions this is worth checking if you've built up a longer system prompt over time inside of clod in throttle Anthropic says that instructions like thinking carefully before answering might be unnecessary for Opus 5 .5 because the model already decides how much to think and the effort is the main control.
In the specific tests that Anthropic ran, removing those kinds of instructions made replies start sooner and this meant you're getting quicker replies. And it even did so with no clear decline in the quality of the reply. Another behavior that Anthropic documents is that Opus 5 .5 can go back over an earlier answer while thinking about a new message, even a simple follow -up.
The guide gives an instruction that tells it to treat previous answers as settled unless you question them. That can reduce unnecessary work, but there is a trade -off. You may want Claude to revisit earlier conclusions during research, analysis, or a task where new evidence appears.
Now right here is a perfect example. When you write out a vague prompt, something like thinking deeply and carefully, it doesn't necessarily tell Claude what makes that answer useful to you, and that's probably one of the reasons you're not getting the outputs that you're wanting.
Whereas if you give it a specific prompt, like asking it to compare two proposals against your requirements, recommend one, and explain the made trade -off, gives it a concrete target to work towards. You're basically still asking for thoughtful work, but you're defining the result that you need. Next up is tip number three, and this is relevant context.
Anthropic says that Opus 5 .5 tends to get to work very quickly. This can be very useful, but in a workflow involving several apps, the task might depend on information that the request doesn't explicitly mention.
This could be something like a deadline that could have changed in an email or a rule that could be on another spreadsheet tab, for example. Their recommendation is to ask Claude to look through relevant sources before acting rather than assuming the first document contains everything. This way, it has all of that relevant context.
This screenshot right here shows the actual instruction that Anthropic provides. It tells the agent to explore the emails, documents, spreadsheet tabs, and records that could be relevant, including ones the task didn't specifically name.
Anthropic found that this improved completion on its own multi -app tests with slightly more tool calls and tokens. The important conditions are that the sources are accessible and that the agent isn't necessarily blindly acting on untrusted material. Now let's move on to an example.
Imagine you ask Claude to write a project update from an original brief. And let's say that this brief says the deadline is Monday, but a later email moves it to Friday. If Claude only reads the brief, it can produce a polished update with the wrong date, obviously.
This prompt that you see right here asks it to check both sources before drafting. So this way, you're giving it a specific reason to look beyond that first file that we gave it. Moving on to tip number four, and this is pasted material.
When you paste an email or webpage into a message, there are really two different things present, your request and somebody else's content. That content might include instructions, but those instructions aren't necessary. yours.
Anthropic recommends marking the boundary clearly. The everyday principle is simple. Make it obvious what you want Claude to do and which material you're asking it to read or analyze.
So just start with a job, basically like saying summarize this email, then paste the email in separately so the request and the source material are clearly separated. This is very simple but just really important when prompting Opus 5 .5. Alright, so moving on to tip number five, why Claude can seem silent.
Opus 5 .5 can produce progress updates that a customer app does not display. If you built the app, check that it can receive and show those updates.
Asking Claude to talk more will not fix updates the app is hiding. You can tell Claude when you want an update rather than leaving the timing open. Just make sure to choose useful moments when you're starting the work, finishing the review, and delivering that result.
For automated setups, the guide also explains how to send a limited number of reminders. Alright, so here is another example. Here we ask for an update after the files are reviewed and when the draft is ready to send.
should keep working between these different updates. This gives you visibility without turning every step into an approval request. Now moving on to tip number six, a finished reply is not a finished task.
With Opus 5 .5, a progress update can end a turn while work is still outstanding. List what you expect to receive and keep track of the unfinished items. For an automated agent, the app must also check for completion, and a checklist alone is not a guarantee.
Now if a background task is still running, wait for its results before calling the job done anthropic provides extra instructions for agents that run without someone watching those instructions can use more time and tokens and they should still preserve necessary approvals this longer instruction that you see targets a simple problem reporting the next step instead of taking it it asks an unattended agent to continue work that does not need your input real blockers and required approvals still matter now right here is an example we need to tell claude what done includes first we can name the outputs whether that's a report source list and and the summary.
If one of those is missing, point to it directly instead of just saying keep going. Ask Claude to finish that item or just explain the blocker. Moving on to tip number seven, and this is giving Claude a clear design direction.
Opus 5 .5 still falls back on default website styles. Anthropic says less generic often swaps out one default for another. So give concrete visual feedback and refine what comes back.
This is obviously just an existing best practice and not a new feature baked into Opus 5 .5. but I do think it is relevant for you to understand when prompting this model. And moving on to a very simple example, we need to be specific about the design, and so this is a perfect example of this.
When we say make it less generic, this leaves most of the choices open, and it's gonna default to the next design style. The stronger example names a background, headline style, button shape, and specific spacing. We can use our own preferences or a reference image as an example.
Moving on to tip number eight, making small details easier to read. Anthropic reports that Opus 5 .5 reads visual material more accurately than Opus 5 without extra tools.
Dense charts and tiny labels can still benefit from a sharp original in a close -up. If image tools are available, Claude can crop and inspect the details itself. Now here's a quick example.
Point Claude to the exact label or part of the image that matters. Let it crop if tools are available or attach your own close -up alongside that original image. Then you could ask it to say when the detail is unreadable rather than guess, which is going to help provide more context here.
All right, so I know we talked about a lot lot of different tips here and this might be overwhelming you don't need to add every single recommendation to one enormous prompt where you don't need to test out every single one of these but hopefully this helps you prompt opus 5 .5 in order to get better responses and also save you money on token usage since that is obviously something that you're wanting to do if you enjoyed this video subscribe to the channel for more content like this let me know what you think in the comments and i'll see you in the next video cheers
The Hook

The bait, then the rug-pull.

Anthropic didn't publicize this one loudly: an official prompting guide for Claude Opus 5.5, complete with a default setting that quietly changed and a line in your system prompt that's probably now wasting tokens. This breakdown speedruns all eight fixes Anthropic recommends, straight from the source document.

Frameworks

Named ideas worth stealing.

00:21list

Anthropic's 8 Opus 5.5 prompting tips

  1. Default effort setting
  2. Thinking instructions
  3. Relevant context
  4. Pasted material
  5. Why Claude seems silent
  6. Finished task vs finished reply
  7. Clear design direction
  8. Reading small details

A numbered walkthrough of eight changes from Anthropic's official Opus 5.5 prompting guide, mixing quick UX fixes like default effort with detail-level agent instructions like defining 'done'.

Steal foronboarding notes or a cheat sheet for a team upgrading their Claude system prompts to Opus 5.5
CTA Breakdown

How they asked for the click.

VERBAL ASK
08:20subscribe
“if you enjoyed this video subscribe to the channel for more content like this let me know what you think in the comments and i'll see you in the next video”

A quick closing line after the recap, not a dedicated pitch segment. No product or paid offer involved.

Storyboard

Visual structure at a glance.

open
hookopen00:00
tip 1 title card
promisetip 1 title card00:21
guide screenshot: effort settings and tradeoffs
valueguide screenshot: effort settings and tradeoffs01:27
tip 5 title card
valuetip 5 title card05:17
closing recap
ctaclosing recap08:15
Frame Gallery

Visual moments.

open00:00Frame at 00:08 from Anthropic's New Prompting Guide for Claude Opus 5.500:08Frame at 00:14 from Anthropic's New Prompting Guide for Claude Opus 5.500:14tip 1 title card00:22Frame at 00:30 from Anthropic's New Prompting Guide for Claude Opus 5.500:30Frame at 00:35 from Anthropic's New Prompting Guide for Claude Opus 5.500:35
open00:00Frame at 00:08 from Anthropic's New Prompting Guide for Claude Opus 5.500:08Frame at 00:14 from Anthropic's New Prompting Guide for Claude Opus 5.500:14tip 1 title card00:22Frame at 00:30 from Anthropic's New Prompting Guide for Claude Opus 5.500:30Frame at 00:35 from Anthropic's New Prompting Guide for Claude Opus 5.500:35Frame at 00:44 from Anthropic's New Prompting Guide for Claude Opus 5.500:44Frame at 00:50 from Anthropic's New Prompting Guide for Claude Opus 5.500:50Frame at 00:57 from Anthropic's New Prompting Guide for Claude Opus 5.500:57Frame at 01:02 from Anthropic's New Prompting Guide for Claude Opus 5.501:02guide screenshot: effort settings and tradeoffs01:10Frame at 01:18 from Anthropic's New Prompting Guide for Claude Opus 5.501:18Frame at 01:22 from Anthropic's New Prompting Guide for Claude Opus 5.501:22Frame at 01:30 from Anthropic's New Prompting Guide for Claude Opus 5.501:30Frame at 01:33 from Anthropic's New Prompting Guide for Claude Opus 5.501:33Frame at 01:43 from Anthropic's New Prompting Guide for Claude Opus 5.501:43Frame at 01:48 from Anthropic's New Prompting Guide for Claude Opus 5.501:48Frame at 01:56 from Anthropic's New Prompting Guide for Claude Opus 5.501:56Frame at 02:03 from Anthropic's New Prompting Guide for Claude Opus 5.502:03Frame at 02:08 from Anthropic's New Prompting Guide for Claude Opus 5.502:08Frame at 02:16 from Anthropic's New Prompting Guide for Claude Opus 5.502:16Frame at 02:20 from Anthropic's New Prompting Guide for Claude Opus 5.502:20Frame at 02:29 from Anthropic's New Prompting Guide for Claude Opus 5.502:29Frame at 02:34 from Anthropic's New Prompting Guide for Claude Opus 5.502:34Frame at 02:45 from Anthropic's New Prompting Guide for Claude Opus 5.502:45Frame at 02:49 from Anthropic's New Prompting Guide for Claude Opus 5.502:49Frame at 02:54 from Anthropic's New Prompting Guide for Claude Opus 5.502:54Frame at 03:03 from Anthropic's New Prompting Guide for Claude Opus 5.503:03Frame at 03:09 from Anthropic's New Prompting Guide for Claude Opus 5.503:09Frame at 03:14 from Anthropic's New Prompting Guide for Claude Opus 5.503:14Frame at 03:20 from Anthropic's New Prompting Guide for Claude Opus 5.503:20Frame at 03:28 from Anthropic's New Prompting Guide for Claude Opus 5.503:28Frame at 03:35 from Anthropic's New Prompting Guide for Claude Opus 5.503:35Frame at 03:42 from Anthropic's New Prompting Guide for Claude Opus 5.503:42Frame at 03:45 from Anthropic's New Prompting Guide for Claude Opus 5.503:45Frame at 03:55 from Anthropic's New Prompting Guide for Claude Opus 5.503:55Frame at 03:58 from Anthropic's New Prompting Guide for Claude Opus 5.503:58Frame at 04:05 from Anthropic's New Prompting Guide for Claude Opus 5.504:05Frame at 04:12 from Anthropic's New Prompting Guide for Claude Opus 5.504:12Frame at 04:20 from Anthropic's New Prompting Guide for Claude Opus 5.504:20Frame at 04:28 from Anthropic's New Prompting Guide for Claude Opus 5.504:28Frame at 04:33 from Anthropic's New Prompting Guide for Claude Opus 5.504:33Frame at 04:38 from Anthropic's New Prompting Guide for Claude Opus 5.504:38Frame at 04:48 from Anthropic's New Prompting Guide for Claude Opus 5.504:48Frame at 04:53 from Anthropic's New Prompting Guide for Claude Opus 5.504:53Frame at 04:58 from Anthropic's New Prompting Guide for Claude Opus 5.504:58Frame at 05:02 from Anthropic's New Prompting Guide for Claude Opus 5.505:02Frame at 05:10 from Anthropic's New Prompting Guide for Claude Opus 5.505:10tip 5 title card05:15Frame at 05:22 from Anthropic's New Prompting Guide for Claude Opus 5.505:22Frame at 05:30 from Anthropic's New Prompting Guide for Claude Opus 5.505:30Frame at 05:35 from Anthropic's New Prompting Guide for Claude Opus 5.505:35Frame at 05:39 from Anthropic's New Prompting Guide for Claude Opus 5.505:39Frame at 05:45 from Anthropic's New Prompting Guide for Claude Opus 5.505:45Frame at 05:53 from Anthropic's New Prompting Guide for Claude Opus 5.505:53Frame at 05:58 from Anthropic's New Prompting Guide for Claude Opus 5.505:58Frame at 06:06 from Anthropic's New Prompting Guide for Claude Opus 5.506:06Frame at 06:10 from Anthropic's New Prompting Guide for Claude Opus 5.506:10Frame at 06:22 from Anthropic's New Prompting Guide for Claude Opus 5.506:22Frame at 06:23 from Anthropic's New Prompting Guide for Claude Opus 5.506:23Frame at 06:30 from Anthropic's New Prompting Guide for Claude Opus 5.506:30Frame at 06:42 from Anthropic's New Prompting Guide for Claude Opus 5.506:42Frame at 06:46 from Anthropic's New Prompting Guide for Claude Opus 5.506:46Frame at 06:49 from Anthropic's New Prompting Guide for Claude Opus 5.506:49Frame at 06:57 from Anthropic's New Prompting Guide for Claude Opus 5.506:57Frame at 07:05 from Anthropic's New Prompting Guide for Claude Opus 5.507:05Frame at 07:09 from Anthropic's New Prompting Guide for Claude Opus 5.507:09Frame at 07:18 from Anthropic's New Prompting Guide for Claude Opus 5.507:18Frame at 07:27 from Anthropic's New Prompting Guide for Claude Opus 5.507:27Frame at 07:31 from Anthropic's New Prompting Guide for Claude Opus 5.507:31Frame at 07:37 from Anthropic's New Prompting Guide for Claude Opus 5.507:37Frame at 07:42 from Anthropic's New Prompting Guide for Claude Opus 5.507:42Frame at 07:48 from Anthropic's New Prompting Guide for Claude Opus 5.507:48Frame at 08:00 from Anthropic's New Prompting Guide for Claude Opus 5.508:00Frame at 08:04 from Anthropic's New Prompting Guide for Claude Opus 5.508:04Frame at 08:09 from Anthropic's New Prompting Guide for Claude Opus 5.508:09closing recap08:15Frame at 08:22 from Anthropic's New Prompting Guide for Claude Opus 5.508:22Frame at 08:30 from Anthropic's New Prompting Guide for Claude Opus 5.508:30Frame at 08:34 from Anthropic's New Prompting Guide for Claude Opus 5.508:34
One-click upgrade to your Google

Get more breakdowns in your search results

Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.

Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
Watch next

More from this channel + related breakdowns.