Anthropic Releases Claude Fable 5: Mythos-Class AI Now Available With Enterprise Safeguards

Anthropic has released Claude Fable 5, making Mythos-class artificial intelligence available to enterprise customers and paid subscribers for the first time. The announcement, made on 9 June 2026, ends two months of restricted access that began when the company unveiled Claude Mythos Preview exclusively to a limited group of cybersecurity partners under its Project Glasswing initiative.

Fable 5 and Mythos 5 share the same underlying model. What distinguishes them is a new layer of safety classifiers that Anthropic has built specifically to enable a broad release of technology that, in unrestricted form, it considers too capable to deploy without controls. Where Mythos 5 retains full capabilities and remains accessible only to vetted partners, Fable 5 routes queries touching cybersecurity, biology, chemistry, and attempts at model distillation to Claude Opus 4.8 rather than responding directly. Anthropic reports that over 95% of Fable 5 sessions trigger no fallback at all.

The Enterprise Implications

On third-party benchmarks measuring software engineering, knowledge work, and vision, Fable 5 outperforms all previous Claude models. Stripe reported that the model compressed months of engineering effort into days, completing a codebase-wide migration across 50 million lines of Ruby in a single day. On Hebbia's Finance Benchmark for senior-level reasoning, Fable 5 recorded the highest score of any model tested, with gains in document reasoning, chart interpretation, and problem-solving.

For GCC enterprises evaluating frontier AI platforms, these are meaningful indicators. The region's largest organisations, particularly those in financial services, government technology, and critical infrastructure, are in the midst of evaluating which AI models can handle long-running autonomous workloads at production scale. Fable 5 is positioned precisely at that threshold. Its ability to sustain complex tasks across millions of tokens, and to apply persistent memory across agentic sessions, addresses operational demands that earlier models could not reliably meet. How the leading frontier models compare for UAE and Saudi enterprise deployments is covered in depth in the GPT-5 vs Claude vs Gemini enterprise platform analysis on AI Watch MENA.

Pricing and Availability

Pricing is set at $10 per million input tokens and $50 per million output tokens, which is double the cost of Claude Opus 4.8 but significantly below the price of Mythos Preview. Anthropic has made Fable 5 immediately available via the Claude API and consumption-based Enterprise plans, with a staged rollout to Pro, Max, Team, and seat-based Enterprise subscription plans running through 22 June 2026.

Pre-IPO Positioning

The launch also advances Anthropic's pre-IPO positioning. The company confidentially filed its IPO prospectus with the US Securities and Exchange Commission days before this release, reporting a revenue run rate of $47 billion and a recent funding valuation approaching $965 billion. The Fable 5 release gives prospective investors a commercially available product that demonstrates the company's ability to monetise Mythos-class capabilities without waiting for the broader regulatory and safety frameworks around unrestricted deployment to mature.

Mythos 5 and Project Glasswing

Separately, Claude Mythos 5 is being made available as an upgrade for all current Project Glasswing partners, including the cybersecurity organisations and critical infrastructure providers that have held access since April. Anthropic has indicated it intends to expand Mythos 5 access through a structured trusted access programme covering both cybersecurity and select biomedical research organisations, with expansion occurring in consultation with the US government. For GCC security leaders, the UAE NESA Cybersecurity Framework sets out the compliance baseline within which enterprise adopters of Fable 5 and Glasswing partners will be operating, while the cybersecurity implications of Mythos-class AI for regional financial infrastructure are examined in the UAE Government Cybersecurity Summit 2026 analysis on MENA Cyber Wire.

The broader signal for enterprise AI buyers is that Mythos-class capability has moved from a restricted preview into a commercially deployed product. GCC technology leaders evaluating AI platform strategy for the second half of 2026 should treat Fable 5 as a material change in what frontier AI can deliver at scale, with the caveat that its most advanced cybersecurity capabilities remain gated behind the Project Glasswing framework.