// news · frontier-models2026-06-11source: anthropic / cnbc / techcrunch

Anthropic releases Mythos 5 to trusted-access partners and Claude Fable 5 to the public — first time the Mythos family ships to paying subscribers, at $10/$50 per million tokens

Anthropic shipped Claude Fable 5 — a Mythos-class model — to Pro, Max, Team, and Enterprise subscribers on June 9, with Mythos 5 going to Project Glasswing and approved trusted-access organizations the same day. Fable is included in paid plans through June 22; on June 23 it shifts to a credit-burn model. Pricing for both lands at $10 input / $50 output per million tokens — double Opus 4.8.

This is the first time the Mythos branch — which Anthropic introduced in April with restricted rollout citing advanced cybersecurity capability — reaches paying retail subscribers in any form. Fable 5 is positioned as the safety-hardened public sibling of Mythos 5: the same underlying architecture with new safeguards that block responses in specific high-risk areas, allowing public distribution where Mythos itself still requires approved-access screening.

The pricing tier signals positioning. Fable 5 at $10/$50 sits a clear step above Opus 4.8's $5/$25 and right at the premium ceiling Anthropic has tested with enterprise buyers. The free-through-June-22 window inside Pro/Max/Team/Enterprise looks like a calibrated trial: Anthropic gets two weeks of paying-subscriber telemetry on Fable's deployment patterns before the credit-burn meter starts. Mythos 5 to Glasswing — JPMorgan, MUFG, SMBC, Mizuho, AWS, Apple, Cisco, Google, Microsoft — continues the segmentation Anthropic has built around capability-tier access programs.

See our analysis →

Anthropic — Claude Fable 5 and Claude Mythos 5 → · CNBC — Anthropic releases Mythos-like AI model to the public, Claude Fable 5 → · TechCrunch — Anthropic releases Claude Fable, a version of Mythos, days after warning AI is becoming too dangerous →