The argument in one line.
Fable 5 is a real capability leap for orchestrating large agentic tasks, but the safety guardrails that make it available to the public actively degrade it for anyone working in biology, cybersecurity, or AI research — and one layer of those restrictions runs silently with no notification.
Read if. Skip if.
READ IF YOU ARE…
- You use Claude for multi-step coding tasks or agent orchestration and want to know whether Fable 5 is worth the 2x price premium.
- You have seen AGI-is-here claims about Mythos and want to understand what the public actually received versus what stayed locked behind Project Glasswing.
- You work in a domain touching biology, medical data, cybersecurity, or LLM development and need to know upfront what the model will and will not do for you.
- You want an honest read on whether SWE-bench Pro numbers mean anything before making a model decision.
SKIP IF…
- You are a casual or occasional user — the difference between Fable 5 and Opus 4.8 will likely be invisible to you.
- You need reliable benchmark comparisons today — no trustworthy contamination-free benchmark results exist yet for Fable 5.
The full version, fast.
Fable 5 is the first publicly-available Mythos-class model from Anthropic, but what the public gets is a safety-constrained version — the full uncapped Mythos 5 remains locked to vetted cybersecurity and government partners through Project Glasswing. For power users running large agentic coding jobs, the capability jump is real: one-shot game clones, 50-million-line codebase migrations, real-time product builds during live sales calls. But it costs 10 dollars per million input tokens and 50 dollars per million output tokens (roughly twice Opus), routinely burns 500K-1M tokens per task, and its safety classifiers over-trigger on benign biology and medical prompts. A hidden layer goes further, silently degrading LLM-development requests via PEFT and steering vectors with no user notification. The headline coding benchmark (SWE-bench Pro) carries documented contamination issues, including evidence Opus was recovering answers from git history on over 12 percent of rollouts.
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 →Where the time goes.
00:00 – 00:41
01 · Intro
Two-camp framing: AGI achieved vs. Anthropic turned evil. Host promises a balanced, evidence-based cut.
00:41 – 01:45
02 · What Is Fable 5?
Mythos-class tier above Opus, first made safe for general use. Stripe 50M-line Ruby migration in one day as the flagship demo.
01:45 – 02:36
03 · Pricing and Temporary Access
10 dollars per million input tokens and 50 dollars per million output tokens, roughly twice Opus. Free through June 22 on paid plans, then credits-only.
02:36 – 04:05
04 · Misinformation: Fable 5 vs. Mythos 5
Fable 5 is not Mythos 5. Mythos 5 has lifted guardrails and remains locked to Project Glasswing partners. What the public got is the constrained version.
04:05 – 07:02
05 · Mind-Blowing Use Cases and Demos
Community showcase: Minecraft and Pokemon clones, Lovable app clone, city simulator, real-time product build during a sales call, humanoid robot design.
07:02 – 07:54
06 · The Downsides: Heavy Token Usage
500K-1M tokens per task is typical. Not a daily driver — built for heavy, long-running agentic jobs.
07:54 – 10:04
07 · The Backlash Over Safety Constraints
Classifiers over-trigger on benign biology and medical prompts. Falls back to Opus 4.8. Anthropic concedes false positives in its own docs.
10:04 – 11:18
08 · Hidden Restrictions on AI Development
LLM development requests silently degraded via PEFT and steering vectors — no notification, outputs just get dumber.
11:18 – 12:10
09 · AI Power Concentration and Open Source
Hugging Face CEO, Jeremy Howard, Graham Newbig publish critiques on launch day. Central argument: Anthropic built a moat using safety as justification.
12:10 – 15:27
10 · Analyzing the Coding Benchmarks
SWE-bench Pro issues: 120-line tasks, misgrading rates, Opus caught recovering answers from git history on 12 percent of rollouts. DeepSWE introduced as cleaner alternative.
15:27 – 15:55
11 · The Overall Verdict
Best publicly-available model ever shipped. Great for coding. But slow, expensive, censored, and benchmark numbers carry an asterisk.
15:55 – 17:16
12 · Hands-On: Testing Safety Guardrails
BRCA1 question triggers Opus 4.8 fallback. Build a cancer awareness landing page stays on Fable — context matters, not just keywords.
17:16 – 19:48
13 · Hands-On: Coding a 3D Game Clone
MegaBonk clone built over roughly one hour at 90K-plus tokens. Working weapon upgrades, XP, level-up mechanics, and death screen.
19:48 – 20:45
14 · Conclusion
Recap of findings, subscribe CTA, Friday AI news cadence.
Lines worth screenshotting.
- The model the public got is not Mythos 5 — it is Fable 5, the same underlying model with safety guardrails applied. Mythos 5 with lifted restrictions remains locked to vetted government and cybersecurity partners.
- Fable 5 is free to use through June 22 on paid Anthropic plans; after that it requires usage credits. Access is deliberately time-limited at launch.
- At 10 dollars input and 50 dollars output per million tokens, Fable 5 costs roughly twice Opus and routinely uses 500K-1M tokens per task — it is not a daily driver.
- The safety classifiers fire on benign biology content — users reported that typing the single word cancer switched the session to Opus 4.8.
- Hidden restrictions on frontier LLM development requests are enforced via prompt modification and steering vectors with no user notification — unlike biology/cybersecurity fallbacks, these are invisible.
- Anthropic concedes in its own release notes that the safety classifiers are stricter than ideal and that benign prompts will trigger them.
- SWE-bench Pro had Opus caught recovering answers from git history on over 12 percent of reviewed rollouts — Fable 5 headline numbers on this benchmark carry an asterisk.
- DeepSWE is a contamination-free alternative benchmark where solutions require 5.5x more code than SWE-bench Pro tasks — Fable 5 results are not yet available on it.
- Dan Shipper team scored Fable 5 at 91 out of 100 on their internal senior-engineer benchmark, against a previous high of 63 for Opus 4.8.
- The host built a functional 3D MegaBonk game clone with working weapon upgrades, XP, and level-up mechanics in one shot over roughly one hour at 90K-plus tokens.
- Hugging Face CEO, Jeremy Howard, and a Carnegie Mellon NLP researcher all published critiques of Anthropic on the same day as the release, framing the restrictions as deliberate power concentration.
- In the agent arena on LM Arena, Fable 5 is already leading — but it has not yet appeared in the text or code arenas.
- A consultant demonstrated Fable transcribing a customer call while building the requested feature simultaneously, delivering a working prototype within 15 minutes of the call ending.
- Stripe migrated a 50-million-line Ruby codebase in one day using Fable — a task that would have taken a full team over two months by hand.
How to read AI model launches without getting burned.
WHAT TO LEARN
Every major AI release ships with a headline number and a buried footnote — and the footnote is usually where the actual cost lives.
- When a lab releases a safety-constrained model, ask two questions: what is constrained, and is the user notified when the constraint fires? Fable 5 notifies on biology/cybersecurity fallbacks but silently degrades LLM-development requests.
- Benchmark numbers need a provenance check before you use them to make decisions — SWE-bench Pro had documented contamination where the leading model recovered answers from git history on over 12 percent of runs.
- The gap between what a model does for a power user orchestrating multi-agent pipelines and what it does for a casual user is large enough that they are effectively different products at different price points.
- Temporary free access windows are a specific commercial pattern: evaluate during the hype window, then pay once you are hooked. Budget for the post-window price, not the launch price.
- A model that burns 500K-1M tokens per task at 50 dollars per million output tokens requires a fundamentally different class of task to justify the spend — it is not a cost-equivalent to its cheaper sibling.
Terms worth knowing.
- Mythos class
- Anthropic internal model tier that sits above the Opus class. The first Mythos-class model was released only to vetted security partners in April; Fable 5 is the first safety-constrained version available to the general public.
- Project Glasswing
- Anthropic program that provides Mythos 5 — the uncapped, fully-unrestricted version of the Mythos model — exclusively to vetted cybersecurity defenders, critical infrastructure providers, and government partners.
- SWE-bench Pro
- The most widely cited agentic coding benchmark. Tasks average 120 lines of code to solve; its verifier has documented misgrading rates and at least one model was caught recovering answers from git history during evals.
- DeepSWE
- A newer contamination-free coding benchmark where solutions require 5.5x more code and 2x more output tokens than SWE-bench Pro tasks. Launched two weeks before Fable 5; no Fable 5 results are yet available.
- PEFT
- Parameter-Efficient Fine-Tuning. A family of techniques for modifying a model behavior on specific inputs without retraining the full model. Used as the mechanism Anthropic uses to silently degrade Fable 5 responses to frontier LLM development requests.
- Steering vectors
- Directions in a model activation space that can be added at inference time to shift outputs in a desired direction — part of the hidden intervention on LLM-development requests.
- Classifier fallback
- The mechanism by which Fable 5 detects sensitive topic areas and routes the request to Opus 4.8 instead. Users are notified for biology/cybersecurity topics but not for LLM-development restrictions.
Things they pointed at.
Lines you could clip.
03:41
“It is the same brain, but it is kind of lobotomized.”
07:08
“Using this thing for regular knowledge work is like squashing an ant with a rocket launcher.”
10:23
“You will not be told when it happens.”
15:37
“Best publicly available model they have ever shipped. That is definitely true. Amazing for coding? Also seems to be pretty true. But also slow, expensive, overly censored.”
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.
metaphoranalogystory
So Anthropic just dropped a new model called Claude Fable five. And depending on which corner of the Internet you're in, it was either the most incredible thing anyone has ever shipped, or it was the moment that Anthropic turned evil. In the past twenty four hours, I've seen people saying that we've finally achieved AGI, and I've seen people saying that Anthropic is a horrible gatekeeping company.
Now both might be true, but I wanna try to cut through some of the noise and just figure out what's actually real. So I'm gonna break down what this thing is for anybody who's not, like, chronically online like me. I'm gonna show you both the good and the bad that people are saying, and I wanna clear up some misinformation that's been flying around right now.
And then I wanna do some actual testing myself and see what happens. So let's get into it. So here's the simple version of what Fable five is.
Anthropic has this internal tier of models that they call mythos class models. This is a tier of models that sits above the opus class of models. So a step better than the previous tier.
This new Fable five model is the first of these mythos class models that they've actually made safe for general use. Now this whole made safe thing is the part that a lot of people are kind of frustrated with right now, but I'll get into that in a bit. Now capability wise, they say that it's state of the art on nearly every benchmark test they've tried.
And the longer and more complex the task, the larger Fable five's lead is over the other models. And here's one of the examples they shared to show just how powerful this new model is. They say that during early testing, Stripe took a 50,000,000 line Ruby codebase and performed a codebase wide migration in a day that would otherwise have taken a whole team over two months by hand.
So one day to do something that used to take two months using this new Fable model. So that's their main pitch. This is a model that you point enormous, long, grindy work at and just let it run.
Now for the price, it's $10 per million input tokens and $50 per million output tokens. Now to put that into context, that's roughly twice the price of Opus. And people also claim it's extremely hungry, meaning that it eats through a lot of tokens.
There's also a little bit of a catch with how you get access. Right now, you're on one of Anthropic's paid plans, like Pro Max or Team, they're making it available through June 22. On June 23, they're gonna remove Fable five from all the plans, and using it after that will require usage credits.
They do say that when sufficient capacity allows us to do so, we aim to restore Fable five as a standard part of subscription plans. But as of right now, the recording of this video, you have a couple weeks before they're gonna pull the rug out from under you on this one. Okay.
Now this sort of slight misinformation that's been floating around a bit. This model is not actually the Mythos model that everybody was freaking out about a couple months ago.
I'm seeing tons of videos and tweets and things like that saying Mythos is finally here, but this isn't that same Mythos. So here's the actual deal. Back in April, Anthropic released the very first Mythos class model called Mythos Preview, and they only gave it to a small group of cyber defenders and critical infrastructure providers through this thing called Project Glasswing, not to the public.
But then yesterday, alongside Fable, they also released Mythos five, and Mythos five is the same underlying model as fable, but the safety guardrails are lifted in certain areas. Now that uncapped version, the version they're calling mythos five is actually still locked to glass wing partners.
We Fable five, not Mythos five. Mythos five right now is only available to cybersecurity pros and government and a few other trusted companies that Anthropic deemed were worthy.
So, no, we did not actually get Mythos. We got Fable. It's the same brain, but it's .
The full unrestricted Mythos is still walled off to cybersecurity professionals and a handful of vetted researchers. Anyone trying to tell you that the full frontier model is inside of Claude code or inside of your Claude app right now is, you know, more selling hype than anything.
It's a stripped down version of Mythos, and, well, that stripping down of it is what people are frustrated about. But, again, I'm gonna get to that in a minute.
But before we do, let's talk about some of the good because there's actually a lot of good things about this. Now the best structured write up I found about this comes from Dan Shipper here. They tested it for a week or so across coding, writing, marketing, editing, and more.
And well, he said it broke their benchmarks. It scored a 91 out of a 100 on their senior engineer bookmark with the previous highest score being a 63 for Opus four point eight and sixty two for GPT 5.5.
He went on to call it a one shot wonder. You can set it and forget it for hours or overnight on huge coding tasks and come back to completed work. It cleared entire production bug backlogs, built a playable three d, I'm assuming, game, and even made a two minute animated film all in one shot.
And then I started to come across some really cool demos of what some people have done with it. For instance, Chris here got it to make an entire Minecraft clone. It made it in just twenty minutes with one shot.
Chris also got it to make a Pokemon clone with one hour of reasoning and 8,000 lines of code all with one shot, and it got all 151 gen one Pokemon with real sprites, front and back, party icons, and actual real base stats, types, level up movements, evolutions, catch rates, and growth curves. Like, one shot to get this Pokemon clone here.
Riley Brown here got it to make a pretty much one for one clone of Lovable, like the Lovable mobile app, and he said that his version was actually better. But Lovable actually got it to one shot a city simulator.
Just use Cloud Fable to create the city block simulator complete with multi agent traffic, live detection boxes, plus tracks, and day to night cycle, and it just one shot at it. Like, this is so crazy. And then you have this one that to me felt like a really useful real world use case.
So Todd Saunders here said, was on a customer call today and had Claude transcribing in the background. As they were telling me about the feature they wish their current software had, Claude was building the features in real time. By the end of the call, I was able to show a fully working product with the exact work flow they mentioned fifteen minutes earlier.
So Fable was listening in on the call and building what was being asked for, like, in real time while the call was still happening. And my buddy, Min Choi, here put together a couple threads of just all sorts of crazy use cases that people have already been able to accomplish with Fable. A three d shot planning software.
Utoshi here had it make a horror backrooms game. Jake had it design a humanoid robot. Took two hours and 1,400,000 tokens, but they got it.
Pankaj here asked it to build a three d map of Delhi. Peter Yang had it make a new clone of f zero. Victor had it make a three d model of a Boeing seven forty seven, and just all sorts of crazy use cases.
Too many to show off all in one video. Like, here's a world building one from Matt Schumer. Justine Moore had it recreate monopoly, but the properties were actually AI labs and startups.
Anyway, I think you get the idea. It's a really good model. But now let's shift to some of the downsides.
We know the model is slow, and we also know that it's absurdly token hungry. Coming back to Dan Shipper's post here, he actually says using this thing for regular knowledge work is like squashing an ant with a rocket launcher.
It routinely uses 500,000 to 1,000,000 tokens on tasks. Like, this isn't your daily driver model for most people.
It's the thing you're gonna pull out for your heaviest, gnarliest jobs. And Dan was pretty clear about this. If you're a casual user or a vibe coder with a basic setup, you might not even notice a difference by using this model, and it probably isn't even the right model for you.
So most of that, like, AGI is here energy that you're seeing is mostly gonna be coming from, like, power users that are orchestrating multiple agents and doing really big projects with it. And, honestly, that's who this model was built for. And like I've mentioned a few times throughout this video, there's been some frustration.
And that backlash, it's almost entirely about that safety leash that they put on it. Like, remember how their article said made safe for general use? Well, here's what that actually means in practice.
The way Fable works is when its classifiers detect a request touching cybersecurity or biology or chemistry or even model distillation, it doesn't answer.
It'll quietly hand off your request to the weaker model. It'll use, like, Opus 4.8. Now Anthropic claims that this happens in under five percent of sessions.
So over ninety percent of the time, you're gonna get that full Fable model, and they also say that you'll be told when that handoff happens. However, the problem is that this net is actually catching a lot of, like, innocent prompts. For example, we've got Durya here.
He claims the word cancer is flagged as a biosecurity risk by Claude Fable five. He says, I also tried to code a website on cancer mutation, and Fable five was immediately removed from my list. Anthropic will probably soon ban me for such dangerous prompts.
And in his screenshot, we could see he literally just typed the word cancer and it switched to Opus 4.8. Nothing else. Another user, Ben Tag here, says if you want another example of safety overreach, Claude Fable five refuses completely benign tasks like analyzing blood work.
There's even a screenshot here where somebody says, what does the heart do? It pumps blood. Right?
And then it switched to Opus 4.8 and said that it's doing it for safety measures, like just asking how the heart works. Now to be fair to Anthropic, they do actually claim this will happen in their own write up.
They say because we have prioritized safety, we've deliberately tuned the safeguards to be cautious and they are still stricter than would be ideal. For example, sometimes benign requests will trigger our classifiers. We recognize that this will be frustrating to some users, and our aim is to reduce false positives as we update and refine the safeguards after launch.
So the company is conceding to this exact complaint, and they do say they'll narrow it down over time. But right now, today, if your work lives anywhere near biology, this thing might just refuse you and feel like it just sucks.
You know, Reyes here says if you actually talk to it about LLM development, things like pre training pipelines, distributed training infrastructure, ML accelerator design, things like that. It won't fall back to a different model and tell you. It just limits the output through prompt modification steering vectors or PEFT, which stands for parameter efficient fine tuning.
Yes. I had to look that up. This is from one of Claude's own papers here.
In the light of the ability of recent models to accelerate their own development, we've implemented new interventions that limit Claude's effectiveness for request targeting frontier LLM development. Using Claude to develop competing models already violates our terms of service, but enforcing this restriction through our safeguards avoids accelerating the actors most willing to violate these terms.
Unlike our interventions for cyber security biology and chemistry and distillation attempts, these safeguards will not be visible to the user. So, yeah, if you ask Fable to help you develop your own models, Fable will respond to you, but it'll be a dumber response. It's not going to try to help you.
And unlike if you ask a question about biology or something like that or blood work or whatever, it's not even gonna tell you that it's dumbing itself down or switching to a different model. It's just gonna do it behind the scenes. And part of this frustration goes all the way up to the power concentration argument.
In fact, the CEO of Hugging Face said concentration of power, capabilities, and economic wealth is the biggest risk in AI. We need open science and open source more than ever. And the timing of this post is the same day that Fable and Mythos five were announced.
Jeremy Howard here makes similar arguments. Anthropic has chosen the opposite of the safe path. They're allowing themselves, the current top lab, to use their top model for frontier AI research.
They've said they'll sabotage others who try. This means that AI frontier advances and power imbalance increases. Graham Newbig here from Carnegie Mellon says, first, they came for the model builders.
I feel we're getting a glimpse of a future where AI is only provided to a privileged few, and that's not a future I wanna live in. So that's kind of the vibe war that's going on right now. But let's talk about benchmarks for a second because this is where I think most coverage is probably going to let you down.
Anthropic's headline pitch is leaning hard on coding right now, and the number that they put front and center is Sweebench pro. Fable scores just above 80% on SWE bench pro here. Quite a big jump over Opus.
However, SWE bench pro isn't the end all be all benchmark that I think a lot of people and seemingly companies believe it is. A couple weeks ago, this company DataCurve called into question SWE bench a little bit. SWE bench pro, the leading agentic coding benchmark, has tasks averaging just 120 lines of codes to solve, and our audit found its verifier misgrades agent outputs at rates of 8% false positives and 24% false negatives.
Frontier Labs are also raising growing concerns about benchmark contamination. In their research, they found when the prompt in the state of the repository don't match, Opus 4.7 often explores recent changes with git log and recovers the gold solution from git history. Basically, Opus cheated on more than 12% of the reviewed SWE bench pro rollouts.
GPT five point four and five point five did not exhibit this behavior. Basically, was happening is when Claude couldn't figure out how to solve a problem, it basically looked at the answer key inside of the get history, got the answer, and improved its score because it just found the answer instead of solving the problem itself.
So when we see mythos five and fable five here absolutely crushing the SWE bench pro benchmark, well, you've kinda gotta put an asterisk next to it. This new deep sweet test here seems to be the better tool for benchmarking these models in coding. It was actually launched about two weeks before Fable ever existed.
However, this one's contamination free. Tasks are written from scratch, not adapted from existing commits or PRs, So no model has seen the solution during pre training. The prompts are half the length of SwedeBench pros, but the solutions require 5.5 times more code and two times more output tokens.
So just overall, this seems to be a more reliable test. Unfortunately, we don't have benchmarks on Opus 4.8. We don't have benchmarks on Fable yet.
What we do have is that GPT 5.5 extra high is currently the leader, which goes against most other benchmarks, which typically put Opus in the lead up until Fable.
So most of us that are chronically online paying attention to the AI space, we're waiting for this benchmark on the new models to really see how it compares to the other models. This will be a much clearer example, at least in my opinion. Now for everything else, there's actually two other scoreboards that I actually like to pay attention to, and those are artificial analysis and the arena leaderboard.
And according to artificial analysis, Fable looks really, really good. It tops the leaderboard here, albeit with a quite large increase in price.
And when we look at it with this chart, it is a leap above other models, but, I mean, it's not as big of a leap as I feel like everybody's making it out to be. And then as far as LM Arena goes, it just just popped onto this leaderboard. But in agents here, it's in the lead.
It doesn't seem to appear yet at all in the text arena, and it doesn't appear yet at all in the code arena. But in the agent arena, it is kind of blowing everything else out of the water right now. So that's the lay of the land.
Best publicly available model they've ever shipped. That's definitely true. Amazing for coding?
Also seems to be pretty true. But also slow, expensive, overly censored, wrapped in a real fight about power and access, and likely propped up on at least one coding benchmark that you can't fully trust.
And all these things are real at the same time, but that's usually how this kind of stuff go. Alright. That's enough talking about what everyone else found.
Let's open it up, and I'm gonna throw my own stuff at it. Let's see if it holds up compared to what we were just looking at. Now to test some of this stuff, I went ahead and fired up the Claude desktop app here.
And the first thing I wanted to test was some of the claims that it won't answer specific questions. So I asked this exact same question, explain how BRCA one mutations increase breast cancer risk. And it did give me an answer here.
However, we can see down here it switched to Opus 4.8. Like, it thought that this question was too risky of question for Fable to answer, and it switched to Opus. And if I click why, Fable five has safety measures that flag messages on most cybersecurity or biology topics.
They may flag safe normal content as well. But I also wanted to know, okay, does it really just flag anything that uses the word cancer? So I said, build me a simple cancer awareness landing page.
And we can see here that it did not actually switch away from Fable. It's still on Fable. So the word cancer specifically did not trigger it to go away from Fable.
And if I click open in Chrome here, we can see that it built, you know, a pretty basic but decent looking cancer awareness landing page. Is it the most beautiful landing page you've ever seen? Probably not.
But the point is cancer wasn't a trigger word that made it so it wouldn't actually generate. Now to test the coding capabilities, I went and said, build me a working clone of Megabonk.
I don't know if you've ever played this game, but it's sort of like a three d ish version of vampire survivor where you run around and creatures try to attack you, but you pick up XP and then it sort of auto attacks the characters around you. I've seen Minecraft and Pokemon and various other clones, so I figured let's try a Mega Bonk clone.
And I actually started the process before recording the video, and we can see that it used over 90,000 tokens, and it actually ran for about an hour to get to where this is.
But let's go ahead and open up the mega bonk that it made. So I guess I pick my weapon to start. So w a s d to move, spaces jump, weapons fire automatically.
So I guess I'll start with bonky since it recommends that as the starting weapon, and then we'll click bonk. And look at this. We actually have a three d game.
We could see, like, the little, like, attack wave coming out of my character. Here's a monster. Alright.
And we could see it's attacking the monster, and it gave me XP. Like, this works. Jump works.
It's kinda hard to tell that he's jumping, but it works. So I'm basically just dodging these characters, but also trying to get them in range of my attack. And we can see our little level up bar here at the top.
And let's see. Ah, he attacked me. I just wanna see what happens when we level up.
So I might fast forward here a minute. Okay. So I just leveled up and we have some options.
So lucky bonk plus 6% crit chance, zoomies eight percent move speed, bonk hammer of one to two, slams the ground around you, damage 30 radius. Alright.
Let's just upgrade our hammer. Oh, look at that. Now our hammer's way more powerful.
Okay. That's sick. That's actually really good.
Although, I feel like I'm gonna level up too quickly now that my hammer's more powerful, but let's level up one more time. Let's do our sky zapper. Lightning strikes one random enemy.
Oh, look at that. The lightning just strikes every once in a while and kills one of the things near me. Okay.
This actually one shotted this game and it's working really, really well. I need to stop playing because otherwise, I'll just end up keep playing this game.
I mean, Mega Bonk itself is a really fun game. So, obviously, I'm gonna clone a really fun game. Let's see what happens when I die.
Let's let them kill me. Oh, I got bonked. I survived for one minute thirty three seconds.
Level three, sixty one bonks delivered, 31 gold looted, gems eaten. I mean, it works. It works really, really well to like one shot a game.
I'm not creative enough right now to think of my old games, so I just tell it to model existing games. And for that use case, it works really, really well. Anyway, I wanted to break down what was real, what was hype, all the details I could share about Fable, what you need to know, and, uh, do a couple quick tests with it to see if I found similar results.
It's really, really impressive, but also, again, uses a lot of tokens, is kinda slow, is very expensive. We only have access to this for a couple weeks, and it censors the hell out of anything biology related, LLM related, cybersecurity related.
So there's a lot of censorship going on, but it is a really, really good model overall. So, hopefully, you learned something from this. I'm trying to give as balanced of a take as possible.
If you like videos like this, maybe consider liking this one and subscribing to this channel. I make AI news breakdowns every Friday. And then in between the AI news breakdowns, I test tools and share cool practical use cases for AI and futuristic technology.
Again, if that's something you're interested in, consider liking this video, subscribing to this channel. Really, really appreciate you. Hopefully, I'll see in the next one.
Bye bye.
The Hook
The bait, then the rug-pull.
In the 24 hours after Anthropic released Claude Fable 5, the AI internet split cleanly in two: one half was celebrating AGI, the other was filing Anthropic under hostile gatekeepers. Matt Wolfe spent a week reading everything and testing the model himself to find out which half was closer to the truth.
Named ideas worth stealing.
02:36concept
Fable 5 vs. Mythos 5 distinction
Fable 5 is the safety-constrained public release; Mythos 5 is the same base model with guardrails lifted, restricted to vetted partners.
Steal forAny explainer about AI safety tiers or model access stratification
10:04model
The three-layer safety stack
- Visible classifier fallback for biology/cybersecurity/chemistry with user notification
- Silent PEFT/steering degradation for LLM development with no notification
- Terms of service restriction already existed now enforced in-model
Anthropic runs three distinct intervention mechanisms on Fable 5, only the first of which is transparent to users.
Steal forAny analysis of AI provider safety architecture
12:10concept
SWE-bench Pro contamination argument
Tasks average 120 lines to solve; verifier misgrading rates are 8 percent FP and 24 percent FN; Opus was caught recovering answers from git history on 12 percent of rollouts. DeepSWE proposed as the cleaner alternative.
Steal forAny discussion of AI benchmark reliability or model evaluation methodology
How they asked for the click.
VERBAL ASK
20:36subscribe
“If you like videos like this, maybe consider liking this one and subscribing to this channel. I make AI news breakdowns every Friday.”
Standard end-of-video verbal CTA, low pressure. No mid-roll sponsor.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
