Anthropic Expands Claude AI Family with New Fable 5.2 and Mythos 5.2 Models

Claude Fable 5.2 and Claude Mythos 5.2, two new AI models, were launched on November 9th by Anthropic. The move comes as the firm gets ready for what would be a historic initial public offering (IPO) as early as next week.

Anthropic boasted that its most recent offerings, Fable 5.2 and Mythos 5.2, produce outputs 30% quicker than their forerunners. In addition to being 20% cheaper than generation 5.1, the models will be accessible for $8 million input tokens and $40 million output tokens, according to the business.

Anthropic Releases Models Amid CEO’s Call for AI Slowdown

This is the first release of so-called "frontier" models from Anthropic since September, when CEO Dario Amodei requested a halt to the rapid advancement of artificial intelligence. According to Anthropic, the models were subjected to extensive alignment and safety testing. In addition, new assessments and post-launch monitoring tools have been developed by the company in advance of the release. During an interview with a media company, Dianne Penn, Anthropic's head of product management for research, clarified that the corporation has requested pacing, not a pause.

The prospectus' lengthy risk factors section underlined the friction between Anthropic's safety measures and heavy commercial and competitive pressures, which the business acknowledged late last month. Additionally, management will need to persuade investors that Anthropic's strategy is the best way to go as they get ready to launch the IPO roadshow. Previously, the corporation had aimed for a valuation of $2 trillion, but now it is looking for $1 trillion to $1.5 trillion in its offering.

Offerings of New Models

Since Fable 5.2 is so good at complicated, open-ended technical work, Penn claimed that "power users or power use cases" are its main target audience. According to Penn, users should utilise Fable 5.2 for challenging engineering, mathematics, and robotics challenges, rather than other models such as Claude Opus 5.5, which were released by Anthropic in September.

According to Penn, these are just brand-new, innovative products, so they are examples of frontier use. As he pointed out, the company anticipates that it may eventually make its way to the Opus class, but for the time being, it believes that the performance of those frontier use cases is superior. Fable 5.2 will be available to the public on November 9th, according to Anthropic, while Mythos 5.2 will be accessible to approved organisations through the company's Cyber Verification Program and Life Sciences Verification Program.

In April, Anthropic debuted Mythos, a model with superior cybersecurity capabilities, however it was initially only made available to a small number of organisations. Two months later, the business unveiled Fable, a publicly accessible variant of Mythos that included more security measures.