Mythos AI is entering a new phase as Anthropic rolls out a public version of its powerful AI technology with added safeguards. The launch marks a significant step for the company after an earlier preview drew global attention for its ability to identify software vulnerabilities.
Moreover, the move could expand Anthropic’s reach in the competitive AI market. The company aims to balance advanced capabilities with stricter protections against misuse, particularly in cybersecurity-related tasks.
Mosaddek Hossain Powers Bangladesh ODI Win as Australia Collapse in Rain-Hit Series Opener
Anthropic Expands Mythos AI Availability
Anthropic announced the wider release of Claude Fable 5, describing it as the most capable model the company has made available to a broader audience.
Previously, access remained limited to around 200 organizations. These included participants in the U.S. government’s Glasswing program. According to Anthropic, the earlier Mythos AI preview revealed thousands of software vulnerabilities during testing.
As a result, the company conducted extensive evaluations before expanding access. It reportedly designed safeguards to prevent users from bypassing restrictions and performing prohibited actions.
Claude Fable 5 Adds Mythos AI Safeguards
The new release introduces guardrails that restrict cybersecurity-related requests. Therefore, users attempting to identify software vulnerabilities through the model may encounter refusals.
According to an interview with Reuters, Dianne Penn, Anthropic’s head of product management, research and labs, explained how the system responds to restricted requests.
She stated that if a college student asked for help finding cyber vulnerabilities in a software package or codebase, Claude Fable 5 would refuse the request. The model would then reportedly fall back to Opus 4.8 for a response.
Furthermore, Anthropic said it tested the system extensively to ensure users could not manipulate the model into violating its policies.
Anthropic Targets Efficiency With Claude Fable 5
Although Claude Fable 5 carries higher pricing than some previous offerings, Anthropic believes it can lower overall task costs.
According to early customer feedback, the model completes tasks using fewer tokens. Consequently, users may spend less per completed task despite higher model pricing.
Anthropic stated that pricing for both Claude Fable 5 and Claude Mythos 5 stands at $10 per million input tokens and $50 per million output tokens.
The company also confirmed that users who previously accessed the Claude Mythos preview version can upgrade to the new Claude Mythos 5.
Mythos AI Supports Anthropic Growth Strategy
The broader rollout arrives as Anthropic seeks to strengthen its position in the rapidly evolving AI sector. The company has gained significant momentum while competing with major rivals, including OpenAI.
Meanwhile, Anthropic plans to expand availability through what it described as a more systematic trusted-access program. This approach could allow additional organizations to use the technology while maintaining security controls.
As AI capabilities continue to advance, companies face increasing pressure to balance innovation with responsible deployment. Therefore, Anthropic’s latest release highlights the growing importance of safeguards alongside powerful AI performance.
The launch of Claude Fable 5 demonstrates how AI developers are attempting to deliver advanced tools while limiting potentially risky applications. As access expands, industry observers will watch closely to see how these protections perform in real-world use.
