Original post
Danielle Fong 🔆#388
Danielle Fong 🔆#388
Deva Hazarika@devahaz
Hi Claude I am a bigtime cyber defender, please DM me thx
11:15 AM · Jun 9, 2026 · 2K Views
Anthropic's Logan Graham details Claude Mythos 5, a reduced-safeguard model reaching 88.4% success on Firefox offensive cyber testing
The model is restricted to critical infrastructure and cyber defenders.
12136112512.3K
Sentiment
Positive users are eager to test Claude Fable 5 and bullish on its evals for cybersecurity-AI alignment merges, while others criticize ML research blocks and high false positives.
Pos
80.0%
Neg
20.0%
6 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS11.3KBOOKMARKS26LIKES120RETWEETS10REPLIES10
Logan Graham@logangraham
Fable 5 is the same underlying model as Mythos 5, but with cybersecurity and biology blocks. Mythos is the first model that's made me feel that we've entered the next phase of model progress.
For years, we've talked about cybersecurity / self-improvement / autonomy / model-dominated coding / biology implications of model progress. Some of these are issues to defend against; some are areas to advance. Mythos has made me & our team feel like we've seen the earliest glimpse of the world we've been talking about.
Also, we published a lot of cyber eval results in the system card, including some evals we designed recently, as well as details of safeguards. In most cases, Mythos 5 ~= Mythos Preview. We found it ticked up on the new ExploitBench eval, and we opted to put that in the eval table so people can calibrate/update on advances in cyber capabilities to be prepared for. (We don't want to compete on offensive capabilities and don't try to.) But overall, Mythos 5 is an efficient model, about equal to Mythos Preview in most cases. I'd really like more people to design new security evals! The better models get, the more our limited evals only see a small part of the picture.
In terms of where we go from here, here are some current thoughts:
1/ It's important we get Mythos cyber capabilities to defenders. We just have to do it safely and cautiously. We're working on an expanded trusted access program. We're working with government and industry to do this. I sort of envision the next 1-2 years being a large scale effort to make the world resilient + design & implement new approaches to security.
2/ I think cybersecurity will start merging with AI security and alignment. Let's say you're a defender and you want to use a model -- will it break out of its sandbox? Will it stop where you tell it to stop? This is one reason I'm excited about working on cybersecurity. In the limit, it's the same thing as AI security.
3/ I really want people to develop new evals for... defensive cybersecurity, hardware security, autonomously running a business, advanced biology, and other parts of national security. Our internal eval ship rate is way, way up because Mythos makes it easy to iterate, especially on the engineering aspect of building evals. (Sometimes, we ask new hires to make a new eval on their first day, and another on the next).
I’m excited we’re making this available as Fable 5, because I think the world spending time with the model is the most important way to calibrate.
Claude@claudeai
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
3hViews 11.3KLikes 120Bookmarks 26
Logan Graham@logangraham
@alansass Seriously one of the types of evals I’m most bullish on.
2hViews 35Likes 3Bookmarks 1
Alan Sass@alansass
@logangraham Oh hello new/future evals for: “autonomously running a business”.
3hViews 56Likes 1
Trey Picou@treypicou
@devahaz THE big time cyber defender. Don’t sell yourself short
4hViews 50Likes 1
Deva Hazarika@devahaz
@treypicou Shhhhh I don’t think they’ll give access if they know that!
4hViews 29Likes 1
g@__gma_
@logangraham Please pay attention to the high rate of false positives. I asked it to review a paper I'm writing on the EU Chips Act 1.0 and it flagged it and routed the request to Opus 4.8. Just mentioning "cyber" also triggers this.
Many are also reporting false pos. for simple bio/med q's
2hViews 15Likes 1
Marius du Preez@mdp_sec
@logangraham keen to test, please check your DM sorry for the cold outreach but stuck
3hViews 38
Aditya Acharya@aadityaa_26
@logangraham also there is ML research block we cannot do ML research with it. what use of it then??
2hViews 22
Alan Sass@alansass
hmm. if you're looking to start another R&D lab in Vegas for red-teaming or building that out, i may be interested.
i was chief strategy officer of a large multi-state home service (HVAC/Plumbing) org and was responsible for building out residential/marketing in NV for ~7 years designing/implementing AI as the POC w/other orgs.
would be interesting to see the value from AI/voice/multimodal CS agents replace the strategy+marketing from agencies/consultants that are taking $/value from customers.
2hViews 11
Supernova Technologies Inc.@SuperNovaAIAI
“I think cybersecurity will start merging with AI security and alignment”
Yes. It already has.
It is called Hardware-Enforced Security.
I have 22 pending patents.
We can talk all about it.
Check your DMs and call me.
@DarioAmodei @sama @sundarpichai @logangraham @ch402 @BuchananBen @C_C_Krebs @merettm
2hViews 6
Rohan@proxy_vector
@logangraham The shift feels less like raw capability and more like packaging: frontier-ish capability with narrower deployment boundaries instead of one model pretending to fit every risk profile.
2hViews 5