Anthropic has begun releasing a new model called “Myth,” based on the same basic model as Mythos, its most powerful AI model class.
Anthropic previously said it developed a model called “Mythos,” which is a modern model of security risks for companies around the world.
At the time, Anthropic noted that Mythos was powerful enough to help bad actors attack public and private software.
“The benefit will be to the side that can benefit the most from these tools,” Anthropic warned in April when it announced the Mythos model.
“In the short term, this could be the attackers, if the border labs are not careful how they release these models. In the long term, we expect that it will be the defenders who will direct resources more effectively and use these models to fix bugs before the new code ever goes live.”
In other words, it could be abused to find and exploit vulnerabilities in applications like Firefox.
Because of those risks, Anthropic decided to limit access to models like the Mythos and only provide them to cyber security professionals and trusted companies.
Now, Anthropic says it has developed strong guard lines for a class of similar models, meaning these powerful AI models can no longer be easily exploited by bad actors.
As a result, it introduced a safer version called “Legend 5.”
Source: BleepingComputer
According to Anthropic, this model has strong defenses in place that will block or deflect sensitive questions, such as those involving cybersecurity, biology, or chemistry, from its previous model, Opus 4.8.
The Claude Mythos 5 is an unrestricted version of that same model, with those defenses raised.
Because of the inherent vulnerability, it is only available to a highly vetted group of trusted partners, such as government cyber defenders and certain life science researchers.
Legend 5 is free for a limited time, and it eats up tokens faster than any other model.
Anthropic says the Fable 5 is an expensive model because it requires more computers, which means the company can’t make it as readily available as the Opus 4.8 or its previous versions.
However, until June 22, Anthropic says that the Fable 5 will be offered to all Pro, Max, and Enterprise customers, but after that period, it will switch to usage-based pricing.
In our tests, BleepingComputer saw the Fable 5 consume a large amount of tokens in a matter of minutes.
Source: BleepingComputer
I noticed this behavior when using Workflow, a new execution system that allows Claude to break complex commands into smaller tasks and wrap similar subagents for execution.
Claude Fable 5 cleared my daily usage of $100 Max subscription, which was zero when I started using it, in just 9 minutes.
Source: BleepingComputer
This doesn’t happen if you casually chat with Claude Fable 5, but if you switch to workflow mode and change the model’s logic to high, you’ll eat all your tokens in minutes.
However, even if you don’t use the Fable 5 with an xhigh effort workflow, you’ll still run it 2 times faster than the Opus model.
This explains why Anthropic is reluctant to open the Fable 5 in the same position as the Opus and other models, but that may change in the coming weeks, as the company is known to touch its models and increase capacity later.
Security teams penetrate 54% of successful attacks and monitor 14%. Some walk around the area without being seen.
The Picus white paper shows how breaches and attack simulations evaluate your SIEM and EDR rules so that threats stop slipping through detection.
Get a white paper
