In a statement, the company indicated that in high-risk fields such as cybersecurity, biology, and chemistry, Fable 5 restrains its own replies and reverts to Claude Opus 4.8. Anthropic conducted extensive stress testing with over 1,000 hours of external red-teaming prior to its launch, finding no universal jailbreaks, though some novel attacks could still be feasible, the statement noted.
Fable 5 is now accessible via the Claude API and consumption-based Enterprise plans. Until June 22, it can be utilized at no additional cost on Pro, Max, Team, and seat-based Enterprise plans. Starting June 23, usage credits will be necessary. Anthropic aims to reintegrate Fable 5 as a standard subscription feature at the earliest opportunity.
Fable 5 and Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens. All data traffic is governed by a 30-day retention policy, even for enterprises with prior zero-retention agreements. Anthropic asserts that data will not be used for training purposes but solely to counter new attacks and minimize false positives.
Alongside this public launch, an expansion of a more discreet program is occurring. Project Glasswing, initiated in April 2026 as a consortium, allows vetted organizations to utilize Mythos for preemptively identifying and addressing software vulnerabilities. The model supporting Fable 5 is the same foundational model as Mythos, now made publicly available with added safety measures.
Initially, Glasswing included around 50 partner organizations, featuring companies like Amazon, Google, Microsoft, Apple, and Nvidia, along with cybersecurity firms like CrowdStrike and Palo Alto Networks. The program collaborates with these organizations to uncover critical zero-day vulnerabilities before they can be exploited by attackers. By the end of May, early adopters had discovered over 10,000 severe security vulnerabilities using the model.
On June 2, Anthropic welcomed about 150 new organizations from over 15 countries into the program, including entities from India. Organizations in India believed to have received or be awaiting access include CERT-In, the National Critical Information Infrastructure Protection Centre, and the Department of Telecommunications’ Digital Intelligence Platform. Notable tech companies, such as Infosys and TCS, have been conducting controlled security assessments, while CERT-In has been reviewing vital systems, including Aadhaar and government login platforms.
This expansion followed weeks of high-level discussions with the government. Finance Minister Nirmala Sitharaman and IT Minister Ashwini Vaishnaw convened to evaluate the risks posed by Mythos to India’s banking sector, leading to directives for the Indian Banks’ Association to create a coordinated response. SEBI also released cybersecurity guidelines for regulated entities, emphasizing immediate patching, AI-driven vulnerability assessments, and zero-trust frameworks. India’s participation signifies a major step in the internationalization of what started as a highly controlled US government initiative.
Meanwhile, the race to go public
The launch of Fable 5 coincides with Anthropic’s preparation for its own IPO. The company confidentially submitted for a US listing on June 1, shortly after completing a $65 billion Series H funding round that valued it close to $1 trillion. OpenAI similarly filed confidentially for its IPO during the same week, and AI chipmaker Cerebras has recently gone public, with SpaceX expected to follow suit next week. The AI class of 2026 is making a simultaneous move to Wall Street.
Last week, Anthropic President Daniela Amodei, speaking at Bloomberg Tech in San Francisco, stated that the transition to public markets is largely driven by the need for capital. “It’s a really big upfront cost to train the models and to serve inference on them,” she explained.
Annualized revenue surpassed $47 billion in May, a significant rise from approximately $9 billion at the end of 2025. Amodei addressed skeptics regarding the effectiveness of AI expenditures, stating that businesses are still in the early phases of discovering how to leverage AI effectively. She highlighted coding, financial services, legal services, and healthcare as industries where value will accumulate over time.
