Anthropic has released two new artificial intelligence models, Claude Fable 5.2 and Claude Mythos 5.2, strengthening its position in the high-end AI market as the company prepares for a potentially historic initial public offering that could take place as soon as next week.
The new models are the company’s most advanced offerings and deliver outputs 30% faster than their predecessors while costing 20% less than the previous 5.1 generation, Anthropic said. The company has set prices at $8 per million input tokens and $40 per million output tokens, bringing greater computational efficiency to products designed for demanding technical applications.
The launch marks Anthropic’s first release of frontier models since Chief Executive Dario Amodei called for a slower pace of AI development in September. The company is attempting to maintain its position in a highly competitive market while demonstrating that more capable systems can be developed and deployed under rigorous safety controls.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
“We called for pacing, not pausing,” Dianne Penn, Anthropic’s head of product management for research, told CNBC.
Anthropic said Fable 5.2 and Mythos 5.2 underwent extensive safety and alignment testing. It also developed new evaluation methods and post-launch monitoring tools before making the models available, an indication that the company is placing greater emphasis on assessing potential risks as its systems become more capable.
The timing places the product launch alongside an important financial milestone. Anthropic is preparing to begin its IPO roadshow, where executives will need to explain how the company’s safety commitments fit with its commercial ambitions and how its technology can sustain growth in a market attracting substantial investment.
The company is reportedly targeting a valuation of $1 trillion to $1.5 trillion for its offering. That range is below the $2 trillion valuation discussed in recent months, suggesting a recalibration of expectations ahead of the listing.
Anthropic is positioning Fable 5.2 primarily for power users and complex technical workloads rather than routine consumer tasks. Penn said the model is designed for open-ended problems involving engineering, mathematics and robotics, where users need more advanced reasoning and problem-solving capabilities.
Other models, including Claude Opus 5.5, released in September, are better suited to everyday tasks, according to Penn. The distinction is important to Anthropic’s product strategy because it allows the company to offer different levels of capability for different workloads rather than relying on a single model for every application.
“These are just net new things that are completely novel, and so they’re frontier use cases,” Penn said. “We expect that over time to maybe trickle to the Opus class, but really right now we do see a difference in terms of performance that we think those frontier use cases benefit from.”
Fable 5.2 became generally available on Monday. Mythos 5.2, however, is restricted to vetted organizations participating in Anthropic’s Life Sciences and Cyber Verification Programs.
That restricted distribution continues the company’s cautious approach to Mythos, which it first introduced in April. Anthropic initially limited access to a select group of organizations because of the model’s advanced cybersecurity capabilities. It subsequently introduced Fable, a publicly available version of Mythos with additional safeguards, two months later.
The decision to maintain restricted access to Mythos 5.2 suggests that Anthropic considers some of its most advanced capabilities too sensitive for unrestricted deployment. Cybersecurity models can help organizations identify vulnerabilities and strengthen their defenses, but similar capabilities could also be misused to discover and exploit weaknesses in computer systems.
However, this could result in a commercial challenge. Limiting access can reduce immediate adoption and revenue opportunities, but unrestricted distribution could increase the risks associated with powerful AI systems. Anthropic must therefore determine which capabilities can be made broadly available and which require additional oversight.
The company said its new evaluations and monitoring tools were developed to support the latest release.
IPO Preparations Put Safety And Revenue Under Scrutiny
Anthropic’s product expansion comes as it prepares to present its business to public-market investors. Its amended prospectus disclosed revenue of $33.24 billion for the nine months ended September 30, representing growth of approximately 1,200% from the corresponding period a year earlier.
That growth provides a substantial commercial foundation for the proposed listing, but it also raises questions about how much of the company’s current performance can be sustained as competition intensifies and the cost of developing advanced models remains high.
Against that backdrop, investors will be assessing more than revenue growth. They will also need to consider the cost of computing infrastructure, the pricing of AI services, customer retention, and the company’s ability to maintain technological advantages as rival developers release competing models.
The new pricing structure offers one indication of Anthropic’s effort to improve the economics of its products. Delivering outputs 30% faster while charging 20% less than the previous generation could make the models more attractive for businesses running complex workloads at scale. Lower prices may encourage adoption, but they can also put pressure on revenue per unit of usage unless customer demand expands sufficiently.
The safety question is equally important to the company’s investment case. Anthropic has made responsible AI development a key part of its public positioning, while acknowledging in the risk factors section of its prospectus that safety obligations and commercial competition can create competing demands.
That tension has got growing attention as the company prepares to market its shares. Investors want evidence that safety controls can keep pace with model capabilities without slowing product development enough to weaken Anthropic’s competitive position.
The release of Fable 5.2 and Mythos 5.2 provides an early indication of how the company intends to manage that balance. It is continuing to develop frontier systems, while using testing, monitoring, and restricted access to manage risks associated with more powerful capabilities.
The proposed IPO will put that approach under closer scrutiny. A valuation of up to $1.5 trillion would require investors to place substantial confidence in Anthropic’s revenue trajectory, technological position and ability to manage the costs and risks of developing advanced AI.