Mythos-class models are soon coming to ALL customers in the COMING WEEKS
Anthropic to release Claude Mythos Preview, a cybersecurity-focused model with Opus-level intelligence, within the coming weeks
The model was developed under Anthropic's Project Glasswing.
Positive users are excited about Anthropic's Mythos-class models rollout for major capability gains and effective safeguards while negative users dismiss the news as marketing hype and predict the models will be heavily restricted.
No Digg Deeper questions have been answered for this story yet.
Most Activity
They are releasing a Mythos-class model with the appropriate safeguards, meaning that you can't use the "too dangerous to release" (mostly cyber) capabilities
Anthropic initially claimed that Mythos was dangerous for every day usage. Now a version of it will be available to everyone.
You are laughing.
We are just a few weeks away from getting access to Mythos and you are laughing.
new from me: anthropic is widely releasing "mythos-class models" in the coming weeks. the company had forecast this last week, but today tucked the timing into the very bottom of a blog post about claude opus 4.8. https://www.bloomberg.com/news/articles/2026-05-28/anthropic-unveils-new-flagship-ai-model-that-s-better-at-coding?srnd=undefined

Like this is just really good and probably gets you a lot of the way to safe deployment for verified users

@ShakeelHashim What are your expectations? I had guessed Opus 4.7 was basically Mythos Preview but deliberately dumbed down on cyber (probably with more general effects), but maybe not?

@scaling01 Have we ever been this back??

@scaling01 they have no other option but to release mythos for everybody
anthropic is upgrading its models way too fast =)

@scaling01 “Safeguards” = More compute

@_AashishReddy I think there’ll be a step change in capabilities

@scaling01 Maybe Mythos will be able to count fingers of a hand and be able to pass the common sense test of the carwash drive/walk.

@curiousiter @scaling01 And when ChatGPT 5.5 started PR with solving cool math stuff

@scaling01 Oh wow they made incredible progress on these safeguards around the same time they acquired more compute to be able to serve mythos (a likely watered down version with the same brand name). How coincidental

@scaling01 mythos with safeguards just means the cyber capabilities that justified the delay are gone so now it's another safe model

@scaling01 "mythos class" is just pure marketing hype

@tunguz Me when i get mythos

@scaling01 I still wonder if they’ll call the model they release Mythos or not

@scaling01 With an introduction of new $500/month subscription Mythos becomes yours !

@scaling01 I guess cybersecurity stocks live to see another day

@scaling01 so that is why they postponed 5.6 release?

@scaling01 Gemini 3.5 Pro is coming in a couple weeks as well. Which means a new GPT is too.