Anthropic introduced two new artificial intelligence models on Monday, Claude Fable 5.2 and Claude Mythos 5.2, as the company moves closer to a highly anticipated initial public offering. The upgrades to the previous 5.1 versions are positioned as faster and more cost-effective, targeting complex technical workloads.
The releases mark Anthropic's first frontier models since CEO Dario Amodei publicly urged a slowdown in AI development. The new models show gains in coding, reasoning, and vision, and are designed for ambitious, open-ended tasks. According to the company, both models run 30% faster than their predecessors and come with lower pricing.
Pricing is set at $8 per million input tokens and $40 per million output tokens in standard mode, with a fast mode available at double the cost. Anthropic highlighted that during prerelease testing, the models helped engineers complete software migrations previously deemed infeasible, design three working electronic devices from scratch, solve two open problems in black hole physics, advance robotics evaluations, and formalize a proof of Fermat's Last Theorem in 30 hours.
Claude Fable 5.2 is available to all users, with Anthropic emphasizing strong safeguards against misuse in cybersecurity and biology. Claude Mythos 5.2 is restricted to organizations pursuing beneficial applications in those fields, subject to verification through Anthropic's life sciences and cybersecurity programs. The company stated that both models underwent extensive safety testing before release.
The launch comes amid shifting expectations for Anthropic's valuation. Earlier hopes of a $2 trillion valuation have faded to around $1.5 trillion, and more recent reports suggest prospective buyers are discussing a range of $1.2 trillion to $1.3 trillion. The company's amended prospectus showed revenue for the first nine months of 2026 surged 1,200% year-over-year, but losses have also widened.
Separately, Anthropic withdrew preliminary IPO pricing details that had been circulated to investment bankers. Initial paperwork proposed selling about 867 million shares at $70 to $80 each, which could have raised up to $69.3 billion. A subsequent email called that information inaccurate, promising an updated filing, though no explanation has been provided.
The withdrawal has left investors uncertain whether it reflects a simple paperwork correction or a broader shift in the company's plans. The move coincides with a broader selloff in AI stocks, which has pressured valuations across the sector. Anthropic's IPO is expected to be one of the largest in recent years, with significant attention from both institutional and retail investors.
Analysts note that the new model releases could bolster Anthropic's competitive position against rivals like OpenAI, but the pricing and performance improvements may also squeeze margins. The company's focus on safety and targeted deployment of its most advanced models could appeal to enterprise clients, though the high cost of development remains a concern.
As the IPO approaches, market participants will be watching for updated financial disclosures and any further changes to the offering structure. The company has not commented on the valuation discussions or the withdrawn pricing details.
This article is for informational purposes only and does not constitute financial advice.
