Anthropic released Claude Fable 5 on June 9, 2026, making it the company’s most capable AI model ever cleared for general public use. The model is the first from Anthropic’s Mythos-class tier – the level that now sits above the Opus class – that the company has cleared for general use. Vellum
Anthropic had previously withheld Claude Mythos from the public, citing concerns that it was too capable at finding high-severity vulnerabilities in major operating systems and web browsers. Anthropic provided the original Claude Mythos Preview only to a small consortium of companies to help secure software systems. Anthropic first made Mythos Preview available in April 2026 through Project Glasswing, a restricted program involving partners including AWS, Microsoft, Apple, and CrowdStrike.
The public release comes as two distinct products. Claude Fable 5 – API model ID claude-fable-5 – is the new generally available flagship, wrapped in safety classifiers that hand certain queries off to Claude Opus 4.8. Alongside it came Claude Mythos 5, the same underlying model with some of its safeguards lifted, available only to a small group of cyber defenders and infrastructure providers. Sensitive cybersecurity, biology, chemistry, and model-distillation requests trigger fallback to Claude Opus 4.8. Anthropic says those safeguards activate in fewer than 5% of sessions.
Anthropic describes Fable 5 as capable of working autonomously for longer than any previous Claude models, with improvements across software engineering, knowledge work, vision, memory, and long-context tasks. On SWE-Bench Pro, a standard software engineering benchmark, the model scores 80.3% – 11 points ahead of the next closest competitor. Fable 5 also leads across knowledge work on the GDPval-AA benchmark at 1932, spatial reasoning at 38.6%, and tool use at 17.4%.
The model supports a 1-million-token context window with a maximum output of 128,000 tokens per request. API pricing is set at $10 per million input tokens and $50 per million output tokens, which is twice the cost of both Claude Opus 4.8 and GPT-5.5’s input price. Converted to Philippine peso at current exchange rates, that puts input costs at roughly ₱570 per million tokens – well above the previous Opus 4.8 tier, though still within reach for developers building on the Claude API.
Filipino developers and businesses using Claude through Anthropic’s API can access Fable 5 today using the claude-fable-5 model string. Anthropic included it in Pro, Max, Team, and seat-based Enterprise subscription plans through June 22, 2026, but said that use after June 23 would require usage credits unless capacity allows it to return as a standard feature. Philippine-based subscribers on the Claude Pro plan can already use the model through the web and mobile apps. Fable 5 and Mythos 5 are designated as covered models with 30-day data retention, which Anthropic says is needed to operate its safety classifiers.
Claude Mythos 5, the unrestricted version, remains limited to existing Glasswing partners. Anthropic had earlier expanded the Glasswing project in early June to 150 organizations across more than 15 countries. Broader access for biology researchers and a wider cybersecurity program is planned, though Anthropic has not committed to a specific timeline.
