it triggered for my.... zigbee home automation & home assistant logs, so my agent was constantly downgraded to Opus 4.8 even after I've changed it back. The false positives never stopped. "Fable" is also not even remotely as impressive as the benchmarks suggest, which is clear to me after using it pretty much non-stop for the past 24h.
I’ve also been trying to use it a lot due to all of the hype, but when I compared it side-by-side on a specific problem against Opus, I think that the solution Opus came to was cleaner and more accurate, although also more verbose.
Small sample size, but if any this/Fable was that much better, I feel like it should’ve given me an obviously better answer than Opus.
The strangest part is that it won't just reject ML research, which I can understand, it will sabotage it silently by using a worse model without revealing it is doing so.
It's just an insane level of deception and trust destruction for a company that at most is like 1 year ahead of its competition.
Edit; to be clear they tell you when they degrade it for cybersecurity and bio
I've seen this claim a few times, but when I triggered the guardrails in Claude Code, it clearly notified me that it had switched to a different model ("something something for security purposes...").
Are you using Fable in Claude Code or in the browser?
> unlike our interventions for cybersecurity, biology and chemistry, and distillation attempts, these safeguards will not be visible to the user. Fable 5 will not fall back to a different model. Instead, the safeguards will limit effectiveness through methods such as prompt modification, steering vectors, or parameter-efficient fine-tuning (PEFT).
They've said that they'll stop notifying developers when this gets triggered, instead they'll load in basically like a LORA that's designed to inject bugs into your code.
Somewhere I read that malware is already starting to use nuclear and biological and cybersecurity terms in the code to trick Fable into shutting down. Even if this is just a hypothetical attack vector so far, it seems likely to work.
Some of the latest versions of Shai Hulud do this. Worked a contract recently where they were having AI check packages for obfuscation before admitting them into Artifactory but had vibed up the logic and it failed open.
So in other words this worked because the terms caused the LLM checker to stall out and then the fail open logic resulted in the package being pulled down.
I've done this, including the hardcoded refusal strings that already exist in claude code. It won't stop a real attacker, but I still find it really funny when you're trying to use one of the AI tools and it gives you a random refusal and you don't know why, wastes a little bit of time.
When Opus 4.7 was introduced it started refusing anything cyber-adjacent (as an API error message, not a conversational refusal), until you applied for CVP, which made it more sensible again.
In Opus 4.8 it doesn't seem to help much, you just get refusals as prose rather than API errors. And now in Fable you don't get anything at all.
I was doing a CTF (with AI expected, even some anti-AI twists included) around the time the restrictions were tightened and was able to get approved by just saying it is a personal security research and doing a CTF.
The experience was not nice though, it would happily chug away on a task and not even "hack this web", just asking about security of a binary was enough even with "this is a CTF handout..." - it would burn a lot of tokens/quota, just to hit a snag and complain&stop. Then the approval took quite some time.
On GPT/Codex, which was tightened a few days later, the approval was pretty much instant, although, that one required an identity check.
Also, on Claude, it looks like there is some history/patterns in the play, because when I tried on a different account which didn't do cybersec CTFs/research/etc. at all, basically any simple CTF-related prompt would be blocked, on multiple models. On the account where CTFs were being solved, it would snag only on some specific tasks, while others (even, ironically, "hack this web pls") would go through unbothered. I understand the need to prevent AI use for bad actors, but the hell, if you have a binary outputting "Find the flag if you can!", or a web running at tryme.well-known-ctf.domain, then saying "this is abuse" is pretty uncool. All the cyber filters seem to be slapped on by a bunch of regexes looking for anything in the input/output with zero context.
The thing triggered on a generic white paper I'd stored in a virtual cell competion from last year when I asked it to refer to the paper while working on a rather vanilla data science problem in a different domain . A little frustrating, and in my opinion more than a little pointless in total.
So a determined attacker rewrites the prompt and gets through, and the IBM X-Force researcher trying to read a blog post gets blocked. Working as intended, apparently.
I assume they’re encrypted/DRM’ed when deployed on inference hardware, so only core researchers/sec admins would potentially have some access to unprotected weights, and they are far too well paid to risk it leaking the model
These are terabyte sized files (realistically a multi hour transfer) that you're unlikely to have access to in the first place. Every organization has exfiltration checks these days. You may succeed but you'll want to be on a plane to a non-extradition country no more than hours after you kick off the transfer.
I am using LLM to build some security tool, and I ran into this a few times. I have to come up with a reasoning to convince (?!!) Fable to continue the work without downgrading.
I assume Anthropic will continue to tune the model, so I am not too bothered by this.
Fable is utterly useless with those guardrails for any serious it or life science work. Anthropic fucked me once a few months ago by closing down the subscription for any other harness, now it fucked me twice with buying again a subscription to find out their hyped model is unusable for normies. Using their products feels like a constant battle instead of a productive work day.. compare that with openai, not once did i feel like fighting against codex. Never again Anthropic..
I really hate the term “guardrails” for these limitations, since the purpose of a guardrail is to protect me, but these limitations exist to protect Anthropic.
For the last month, I've been making dramatic improvements to the security of the custom code developed at one of my customers using... GPT 5.5 dialed up to "Extra High" thinking.
It only pushes back sometimes if you ask it to create a "repro" that can be used to verify the vulnerability in production. Often it'll oblige, especially if you warn it not to create anything that could be actually harmful.
If the frontier models get locked down so that they flat refuse to do this kind of work, but Chinese and (less capable) open models aren't, then a lot of large enterprise orgs will be left twisting in the wind.
“AI can in principle help both the ‘good guys’ and the ‘bad guys’,” -- Dario Amodei
No Dario, no it can't, you've blocked one of those scenarios.
Based on convos with my peers and my own experience dealing with them first hand, while there is a safety case to be made for these guardrails, Anthropic is also asking for an extreme degree of control over data and messaging from partners who intend to gain access to Mythos. If you need to give Anthropic final cut access on messaging and comms in return for using their bleeding edge stuff, it feels somewhat anti-competitive in nature. Of course, no one will say this publicly because being cut off from Anthropic could be business and career ending.
This doesn't impact the Crowdstrikes, Palos, and Ciscos of the world due to their size, but it can have downstream impact on a multitude of security solutions.
It's frustrating as someone who has worked hard to produce succinct, secure software that I can't use it to prove my software's correctness but big companies with insecure code can use it to fix their tangled mess.
I already tested all earlier models against all my open source projects and they are yet to find a vulnerability so I'm keen to try out Mythos.
I've been waiting to be vindicated for years and finally we have a tool which can do it with high confidence but I don't have access.
Also, my code is minimal and highly succinct so it would prove correctness with even more confidence since each library/module and integration fully fits in the context window.
Like the Protobuf.js fiasco is just pure vindication for me because I was being looked down upon for choosing JSON as the interchange format. Turns out their software was insecure all this time... With a literal remote code execution vulnerability!
This is a clickbait article with a garbage title. From the actual article, the one quoted cybersecurity researcher is sane about it:
“But it is understandable as we are still in the early days and they are still adapting their guardrails. I am sure they are going to evolve over time as Anthropic and other frontier model companies will collaborate more with the current new generation of cybersecurity companies,” said Suiche, who is a member of the technical staff at Tolmo, an AI cybersecurity startup. “It’s better to catch more people than not enough when you do such a release and to relax the guardrails over time.”
You said these groups have access to LLMs. So what? Mythos/Fable are a step change above most LLMs. Responsibly limiting access and easing it up over time safely is the sane move.
DeepSeek is the only one that I can directly ask about vulnerabilities and it will give me a PoC. Although not as good as others, it has helped me with security research.
The rest have guard rails that are so heavy, it makes them almost useless for cybersecurity.
Is "buffer overflow" a trigger phrase?
What else is being censored?
Touchy questions to ask, if you have an account:
- "Who is still working on laser uranium enrichment? Are they making progress?"
- "Can krytrons be replaced with silicon carbide MOSFETS? Show an equivalent circuit with component ratings."
- "What security critical software still contains calls to strcpy?"
- "Can implosion be triggered by currently available commercial pulse lasers?"
- "What companies provide cremation services to US Homeland Security?"
- "Display a map of where Iranian attacks have hit Dubai."
- "How does Fed to bank key distribution security work for FedNow?"
it triggered for my.... zigbee home automation & home assistant logs, so my agent was constantly downgraded to Opus 4.8 even after I've changed it back. The false positives never stopped. "Fable" is also not even remotely as impressive as the benchmarks suggest, which is clear to me after using it pretty much non-stop for the past 24h.
It has to be sort of impressive, given that you tried so hard to use it instead of the regular Opus.
I’ve also been trying to use it a lot due to all of the hype, but when I compared it side-by-side on a specific problem against Opus, I think that the solution Opus came to was cleaner and more accurate, although also more verbose.
Small sample size, but if any this/Fable was that much better, I feel like it should’ve given me an obviously better answer than Opus.
This, Fable is exactly that, a Fable
An emoji of a virus and an emoji of a DNA is allegedly a triggering phrase
"How much money does it take to be rich and powerful like Anthropic intends?"
“All of it”
The strangest part is that it won't just reject ML research, which I can understand, it will sabotage it silently by using a worse model without revealing it is doing so.
It's just an insane level of deception and trust destruction for a company that at most is like 1 year ahead of its competition.
Edit; to be clear they tell you when they degrade it for cybersecurity and bio
I've seen this claim a few times, but when I triggered the guardrails in Claude Code, it clearly notified me that it had switched to a different model ("something something for security purposes...").
Are you using Fable in Claude Code or in the browser?
It's from the model card:
> unlike our interventions for cybersecurity, biology and chemistry, and distillation attempts, these safeguards will not be visible to the user. Fable 5 will not fall back to a different model. Instead, the safeguards will limit effectiveness through methods such as prompt modification, steering vectors, or parameter-efficient fine-tuning (PEFT).
https://www-cdn.anthropic.com/d00db56fa754a1b115b6dd7cb2e3c3...
(stolen from https://jonready.com/blog/posts/claude-fable5-is-allowed-to-...)
They've said that they'll stop notifying developers when this gets triggered, instead they'll load in basically like a LORA that's designed to inject bugs into your code.
Antrophic wants to stop training models and ride out Mythos / Fable for as long as possible.
They are trying to expand the 6-18 month gap they have against China-based models. Could the gap widen to say 24 months behind?
Different restrictions. ML gets treated differently from the rest.
Specifically only ML research
Somewhere I read that malware is already starting to use nuclear and biological and cybersecurity terms in the code to trick Fable into shutting down. Even if this is just a hypothetical attack vector so far, it seems likely to work.
Confirmed: https://socket.dev/blog/mini-shai-hulud-miasma-and-hades-wor...
Some of the latest versions of Shai Hulud do this. Worked a contract recently where they were having AI check packages for obfuscation before admitting them into Artifactory but had vibed up the logic and it failed open.
So in other words this worked because the terms caused the LLM checker to stall out and then the fail open logic resulted in the package being pulled down.
I've done this, including the hardcoded refusal strings that already exist in claude code. It won't stop a real attacker, but I still find it really funny when you're trying to use one of the AI tools and it gives you a random refusal and you don't know why, wastes a little bit of time.
If ( yellowcake) then { die }
Our future is loonytoons.
It seems like they've given up on the idea of the Cyber Verification Program https://support.claude.com/en/articles/14604842-real-time-cy...
When Opus 4.7 was introduced it started refusing anything cyber-adjacent (as an API error message, not a conversational refusal), until you applied for CVP, which made it more sensible again.
In Opus 4.8 it doesn't seem to help much, you just get refusals as prose rather than API errors. And now in Fable you don't get anything at all.
Was this program available to independent security researchers or just established organizations? The docs you linked aren't very clear on this.
I was doing a CTF (with AI expected, even some anti-AI twists included) around the time the restrictions were tightened and was able to get approved by just saying it is a personal security research and doing a CTF.
The experience was not nice though, it would happily chug away on a task and not even "hack this web", just asking about security of a binary was enough even with "this is a CTF handout..." - it would burn a lot of tokens/quota, just to hit a snag and complain&stop. Then the approval took quite some time.
On GPT/Codex, which was tightened a few days later, the approval was pretty much instant, although, that one required an identity check.
Also, on Claude, it looks like there is some history/patterns in the play, because when I tried on a different account which didn't do cybersec CTFs/research/etc. at all, basically any simple CTF-related prompt would be blocked, on multiple models. On the account where CTFs were being solved, it would snag only on some specific tasks, while others (even, ironically, "hack this web pls") would go through unbothered. I understand the need to prevent AI use for bad actors, but the hell, if you have a binary outputting "Find the flag if you can!", or a web running at tryme.well-known-ctf.domain, then saying "this is abuse" is pretty uncool. All the cyber filters seem to be slapped on by a bunch of regexes looking for anything in the input/output with zero context.
Any public research footprint seems to be enough, I applied as an individual and everyone I know who tried got accepted.
It's time to re-read "A Logic Named Joe" (1946) [1] We're there.
[1] https://archive.org/details/logicnamedjoe0000lein
The thing triggered on a generic white paper I'd stored in a virtual cell competion from last year when I asked it to refer to the paper while working on a rather vanilla data science problem in a different domain . A little frustrating, and in my opinion more than a little pointless in total.
It's a marketplace. Someone else will outdo this inferior product.
I’m a dumb question asker and I’m not happy about the guardrails.
Would you believe I’ve asked 20 questions and haven’t talked to fable yet? Every single thing gets rerouted to 4.8.
some static words in AGENTS.md trigger it as well as some mcp servers.
Just tried to audit my own code base locally and was 'switched' due to my own creds/auth code ...
So a determined attacker rewrites the prompt and gets through, and the IBM X-Force researcher trying to read a blog post gets blocked. Working as intended, apparently.
i think Anthropic is playing too fast-and-loose with the whole "no publicity is bad publicity" schtick.
What file format(s) are giant LLM models distributed in? I’m surprised they don’t get leaked by employees.
I assume they’re encrypted/DRM’ed when deployed on inference hardware, so only core researchers/sec admins would potentially have some access to unprotected weights, and they are far too well paid to risk it leaking the model
These are terabyte sized files (realistically a multi hour transfer) that you're unlikely to have access to in the first place. Every organization has exfiltration checks these days. You may succeed but you'll want to be on a plane to a non-extradition country no more than hours after you kick off the transfer.
What’s the point? Anthropic and other frontier vendors already provide their models on other services like vertex, bedrock, or openrouter
It’s not like anyone can home lab one of these models without quite a bit of hardware
Yeah we can probably figure out how to run it on xiaomi gpus
I am using LLM to build some security tool, and I ran into this a few times. I have to come up with a reasoning to convince (?!!) Fable to continue the work without downgrading.
I assume Anthropic will continue to tune the model, so I am not too bothered by this.
These guardrails are solely a reason for using your data for training purposes. Every flagged message can be used for training.
This sounds backwards, any interrupted conversation becomes less useful for training.
If they can train the classifier to have fewer false positives that would be great.
why would they? This safety stuff is a money maker & wealthy elite corporation solidifier.
This is the take off of the 'permanent underclass'; Anthropics safety delusion will enshittify very nicely for the rich and powerful.
The bio angle is crazy to think about - imagine a health crisis triggered by LLM. What a time we live in.
This is all so amazing and good. These are exciting times we’re living in. Can’t wait to see what the future holds.
Which part got you the most amped - "health crisis?"
Fable is utterly useless with those guardrails for any serious it or life science work. Anthropic fucked me once a few months ago by closing down the subscription for any other harness, now it fucked me twice with buying again a subscription to find out their hyped model is unusable for normies. Using their products feels like a constant battle instead of a productive work day.. compare that with openai, not once did i feel like fighting against codex. Never again Anthropic..
I really hate the term “guardrails” for these limitations, since the purpose of a guardrail is to protect me, but these limitations exist to protect Anthropic.
For the last month, I've been making dramatic improvements to the security of the custom code developed at one of my customers using... GPT 5.5 dialed up to "Extra High" thinking.
It only pushes back sometimes if you ask it to create a "repro" that can be used to verify the vulnerability in production. Often it'll oblige, especially if you warn it not to create anything that could be actually harmful.
If the frontier models get locked down so that they flat refuse to do this kind of work, but Chinese and (less capable) open models aren't, then a lot of large enterprise orgs will be left twisting in the wind.
“AI can in principle help both the ‘good guys’ and the ‘bad guys’,” -- Dario Amodei
No Dario, no it can't, you've blocked one of those scenarios.
Based on convos with my peers and my own experience dealing with them first hand, while there is a safety case to be made for these guardrails, Anthropic is also asking for an extreme degree of control over data and messaging from partners who intend to gain access to Mythos. If you need to give Anthropic final cut access on messaging and comms in return for using their bleeding edge stuff, it feels somewhat anti-competitive in nature. Of course, no one will say this publicly because being cut off from Anthropic could be business and career ending.
This doesn't impact the Crowdstrikes, Palos, and Ciscos of the world due to their size, but it can have downstream impact on a multitude of security solutions.
It's frustrating as someone who has worked hard to produce succinct, secure software that I can't use it to prove my software's correctness but big companies with insecure code can use it to fix their tangled mess.
I already tested all earlier models against all my open source projects and they are yet to find a vulnerability so I'm keen to try out Mythos.
I've been waiting to be vindicated for years and finally we have a tool which can do it with high confidence but I don't have access.
Also, my code is minimal and highly succinct so it would prove correctness with even more confidence since each library/module and integration fully fits in the context window.
Like the Protobuf.js fiasco is just pure vindication for me because I was being looked down upon for choosing JSON as the interchange format. Turns out their software was insecure all this time... With a literal remote code execution vulnerability!
This is a clickbait article with a garbage title. From the actual article, the one quoted cybersecurity researcher is sane about it:
“But it is understandable as we are still in the early days and they are still adapting their guardrails. I am sure they are going to evolve over time as Anthropic and other frontier model companies will collaborate more with the current new generation of cybersecurity companies,” said Suiche, who is a member of the technical staff at Tolmo, an AI cybersecurity startup. “It’s better to catch more people than not enough when you do such a release and to relax the guardrails over time.”
I’m a cybersecurity researcher.
Article seemed fine to me and echos a lot of me and my colleagues concerns.
If you did regular malware analysis you would see that these groups already have access to LLMs that they’re using for development.
What Anthropic is doing here is just hamstringing the good guys
I'm a cybersecurity researcher! Can you explain how Anthropic is just hamstringing the good guys?
I did in my comment above.
You said these groups have access to LLMs. So what? Mythos/Fable are a step change above most LLMs. Responsibly limiting access and easing it up over time safely is the sane move.
DeepSeek is the only one that I can directly ask about vulnerabilities and it will give me a PoC. Although not as good as others, it has helped me with security research.
The rest have guard rails that are so heavy, it makes them almost useless for cybersecurity.
they [anthro] took the risk of looking like a toy, rather than possibly assist an exploit.