Anthropic expands Mythos-class model access with Claude Fable 5

Traders Union
All News
Financial News
Anthropic expands Claude Fable 5 access

Highlights

Anthropic expands Claude Fable 5 access due to new safety measures, routing high-risk prompts to Claude Opus 4.8 and maintaining Mythos-class capability.
Claude Fable 5 outperforms Claude Opus 4.8 by over 10% on key software engineering and knowledge work benchmarks, broadening safe public-facing deployment.
Anthropic files confidential IPO prospectus after reporting a $47 billion revenue run rate in May and closing funding at a $965 billion valuation, surpassing OpenAI's $852 billion.

Safeguards underpin broader launch

As reported by CNBC, Anthropic says the wider release of Claude Fable 5 is possible because it has introduced new safety measures that block responses in specific high-risk categories. The company positions the model as Mythos-class in capability while restricting dangerous outputs and, in some cases, routing users to Claude Opus 4.8 for a safer response.

Dianne Penn, Anthropic's head of product management for research, says the company is trying to win what it calls a "race to the top" by delivering valuable technology with guardrails that increase benefits relative to harm. She says Claude Fable 5 marks a significant jump in capability, prompting Anthropic to build new classifiers and additional protections for the launch.

Anthropic says Claude Fable 5 delivers exceptional performance across software engineering and knowledge work tasks. In a company blog post, it says the model scores more than 10% higher than Claude Opus 4.8 on some benchmarks.

Anthropic also launches Claude Mythos 5 on Tuesday, describing it in a blog post as the same underlying model as Claude Fable 5 but with safeguards lifted in some areas. That keeps the company on track with its stated eventual goal of deploying Mythos-class models at scale while maintaining tighter controls for broader public-facing access.

IPO backdrop sharpens investor focus

The product release arrives days after Anthropic says it confidentially files its IPO prospectus with the Securities and Exchange Commission, setting up what could become a major public offering. The launch also builds on momentum created in April, when Mythos drew attention from Wall Street and government officials for its ability to identify security flaws in software.

At that time, Anthropic says it does not plan to make Mythos generally available and instead limits access to a select group of companies through its cybersecurity initiative Project Glasswing. Claude Fable 5 now signals that the company is prepared to bring an equally powerful model to a much broader customer base under stricter controls.

Anthropic says in May that its revenue run rate reaches $47 billion, up from roughly $10 billion in annual revenue last year. The company also recently closes a funding round at a $965 billion valuation, above chief rival OpenAI's $852 billion valuation in late March, underscoring the commercial and competitive stakes around its latest model rollout.

Our earlier article on SpaceX’s AI compute contracts with Google and Anthropic explained how those multiyear deals could add more than $25 billion in annualized revenue ahead of SpaceX’s planned IPO. We also noted investor concerns about the contracts’ structure, including delivery milestones and relatively short cancellation windows, which could affect how durable that new revenue looks in public markets.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.