Anthropic introduces two advanced AI models with 30% faster performance and 20% lower costs, marking its first frontier model launch since CEO Dario Amodei called for a slowdown in AI development.
Anthropic has unveiled two advanced artificial intelligence models, Claude Fable 5.2 and Claude Mythos 5.2, marking a significant milestone for the AI company as it prepares for a potential initial public offering (IPO) that could take place as early as next week.
The latest releases represent Anthropic's most advanced AI offerings to date. According to the company, both models deliver outputs approximately 30% faster than their previous-generation counterparts while costing 20% less than the Claude 5.1 series.
Anthropic has announced pricing of $8 per million input tokens and $40 per million output tokens, making the new models more cost-efficient for businesses and developers handling complex AI workloads.
Anthropic Resumes Frontier AI Development After Safety Slowdown Call
The launch is particularly significant because it marks Anthropic's first release of frontier AI models since CEO Dario Amodei called for a slower pace of artificial intelligence development in September.
The company has emphasised that responsible AI development remains central to its strategy. Anthropic said Claude Fable 5.2 and Claude Mythos 5.2 underwent extensive safety and alignment evaluations before their release. It also developed additional assessment frameworks and post-launch monitoring systems to identify potential risks.
"We called for pacing, not pausing," Dianne Penn, Anthropic's head of product management for research, told CNBC.
The statement reflects Anthropic's attempt to balance technological innovation with responsible AI development as competition across the global artificial intelligence industry intensifies.
Anthropic's IPO Plans Draw Attention From Investors
The launch comes at a crucial time for Anthropic as the company prepares for a possible public listing. Its IPO preparations are attracting significant attention from investors, particularly as the company works to demonstrate its technological leadership and long-term commercial potential.
However, Anthropic faces the challenge of convincing prospective investors that its commitment to AI safety can coexist with the commercial demands of operating in an increasingly competitive market.
According to previous CNBC reporting, Anthropic is targeting a potential IPO valuation of between $1 trillion and $1.5 trillion. This represents a reduction from the $2 trillion valuation target reportedly discussed in recent months.
The company's prospectus also highlighted several business and operational risks, including challenges associated with rapid AI development, competition and the financial implications of building advanced AI infrastructure.
Claude Mythos and Fable Strengthen Anthropic's AI Position
Anthropic initially introduced Claude Mythos in April, restricting access to a selected group of organisations because of its advanced cybersecurity capabilities.
Two months later, the company introduced Claude Fable as a more widely accessible version of Mythos, incorporating additional safeguards to support broader usage.
Both models attracted considerable attention from Wall Street and government officials, contributing to Anthropic's growing influence in the artificial intelligence market.
The company's financial performance has also reflected its rapid expansion. According to its amended IPO prospectus, Anthropic generated $33.24 billion in revenue during the nine months ended September 30, representing an increase of approximately 1,200% compared with the corresponding period last year.
Claude Fable 5.2 Targets Advanced Technical and Engineering Tasks
Anthropic says Claude Fable 5.2 is designed primarily for advanced users and demanding technical applications.
Dianne Penn explained that the model is particularly effective at complex, open-ended technical assignments, making it suitable for areas such as software engineering, mathematics and robotics.
While Claude Opus 5.5, introduced in September, is positioned for everyday productivity and general-purpose tasks, Fable 5.2 is intended for more challenging problems that require advanced reasoning and technical capabilities.
"These are just net new things that are completely novel, and so they're frontier use cases," Penn said.
She added that some of these advanced capabilities could eventually become available in Anthropic's Opus-class models. For now, however, the company sees a meaningful performance distinction between its frontier models and offerings designed for everyday applications.
Claude Fable 5.2 Available Globally, Mythos 5.2 Restricted
Anthropic confirmed that Claude Fable 5.2 became generally available on Monday, allowing users to access its latest capabilities.
Claude Mythos 5.2, meanwhile, will remain restricted to vetted organisations participating in Anthropic's Life Sciences and Cyber Verification Programs.
The controlled rollout reflects the company's cautious approach to distributing powerful AI capabilities, particularly those that could have significant cybersecurity implications.
With the release of Fable 5.2 and Mythos 5.2, Anthropic is strengthening its position in the frontier AI market while preparing for a potentially landmark IPO. The company's ability to combine advanced AI performance, commercial growth and responsible development will remain central to its next phase of expansion.