Anthropic has launched Claude Fable 5, marking its first Mythos-class model available to the public.

      TL;DR: Anthropic has introduced Claude Fable 5, its first publicly accessible Mythos-class model, which includes safeguards that prevent high-risk queries and defaults to Opus 4.8 for safer responses. Released on Tuesday, Fable 5 utilizes the same architecture as the restricted Mythos system, marking the first public availability of Mythos-class intelligence. It is accessible to enterprise clients and paid users, featuring new safeguards that block responses in areas like cybersecurity and biology, reverting to Claude Opus 4.8 when necessary. Early data shows the fallback occurs in less than 5% of sessions, so most users will interact with Fable 5 directly.

      Dianne Penn, Anthropic’s head of product management for research, stated that their focus is on providing valuable technology while ensuring robust safety measures to maximize benefits over potential harm. The model demonstrates a significant increase in capability compared to Opus 4.8, which is why these additional safeguards were implemented. Fable 5 reportedly scored over 10% higher on some benchmarks than Opus 4.8, but independent assessments are still pending.

      Additionally, Anthropic has launched Claude Mythos 5, an updated version of the restricted Mythos model with the cybersecurity and biology safeguards removed. Mythos 5 remains available exclusively to around 50 vetted partners in Project Glasswing, the company's cybersecurity initiative, which has identified over 10,000 significant vulnerabilities in major software projects since April. Both models share the same foundational architecture, with Fable 5 incorporating the safety measures deemed necessary for broad deployment.

      The pricing reflects this enhancement, with Fable 5 priced at $10 per million input tokens and $50 per million output tokens, which is twice the cost of Opus 4.8, released last month. Until June 22, Fable 5 is included at no extra charge for Pro, Max, Team, and seat-based Enterprise subscribers, after which access will require usage credits.

      Early customers have indicated increased efficiency per task, arguing that Fable 5’s enhanced intelligence reduces the number of iterations required to complete tasks. However, these claims are based on feedback from early users and have not been independently verified.

      The launch is strategically timed for commercial impact, coinciding with Anthropic’s confidential IPO filing with the Securities and Exchange Commission last week, potentially leading to a listing this autumn. The company reported a revenue run rate exceeding $47 billion, a significant increase from approximately $10 billion in annual revenue for 2025, and recently completed a $65 billion Series H funding round, resulting in a $965 billion post-money valuation.

      This valuation positions Anthropic ahead of OpenAI for the first time. OpenAI's last funding round concluded in March at $852 billion, and the company filed its own confidential IPO paperwork recently. SpaceX, which acquired Elon Musk's AI startup xAI, is anticipated to start trading on Friday at a valuation close to $1.8 trillion.

      These developments suggest that three of the most valuable private tech companies are preparing to enter public markets in close succession, with a potential for over $200 billion in new public market value from their listings by the fall of 2026. Claude Fable 5 exemplifies Anthropic's aim to provide advanced intelligence to customers at scale while avoiding the severe misuse risks associated with the original Mythos. The company has also established a 30-day data retention policy for all Fable 5 and Mythos 5 traffic for safety oversight, although it maintains that the data will not be used for training purposes.

      However, it remains unclear whether a model that has a fallback rate of less than 5% can effectively mitigate the types of attacks for which Mythos Preview was designed. For investors considering Anthropic’s IPO potential, the launch of Fable 5 addresses a critical strategic issue: how to monetize Mythos-class capabilities without releasing the cybersecurity toolkit that poses both value and risks. The current solution is a model that is generally more intelligent across the board, at double the cost, and future quarters will reveal if this balance of trade-offs generates sufficient revenue to support a nearly trillion-dollar public valuation.