TL;DR

Claude Mythos is 📝Anthropic's Mythos-class 📝Artificial Intelligence (AI) model. It reached the public on June 9, 2026 as 📝Claude Fable 5 (general availability) and 📝Claude Mythos 5 (restricted), following the earlier Claude Mythos Preview research release.

Overview

Claude Mythos is Anthropic's most capable model tier, positioned above Claude Opus 4.8. It first surfaced in early-2026 leaks, was acknowledged as a restricted research model — Claude Mythos Preview — deployed through Project Glasswing, and reached general availability on June 9, 2026 as the safety-locked Claude Fable 5 alongside the less-restricted Claude Mythos 5. This memo tracks the path from rumor to release and separates what Anthropic has confirmed from what remains reported.

Key Facts

Developer: Anthropic
Tier: Mythos-class — Anthropic's most capable, above Claude Opus 4.8
General availability: June 9, 2026 — Claude Fable 5 (public) and Claude Mythos 5 (restricted)
Fable 5 model ID: claude-fable-5
Pricing (Fable 5 / Mythos 5): $10 / input MTok, $50 / output MTok
Research preview: Claude Mythos Preview, deployed via Project Glasswing to defensive-cybersecurity partners

Sources: Anthropic — Fable 5 & Mythos 5, Anthropic — Project Glasswing.

Timeline

Late 2025 – early 2026 — A model codenamed or branded "Mythos" appears in leaked materials, described as a tier above Claude Opus with substantially higher compute requirements. [reported]
March 2026 — Leaked blog-post drafts make the name public; Anthropic acknowledges the model's development and its cybersecurity risk to reporters. [reported, per press]
Q1–Q2 2026 — Anthropic deploys Claude Mythos Preview through Project Glasswing, a restricted program to secure critical software, with 12 launch partners plus 40+ additional organizations. [confirmed — Anthropic]
June 9, 2026 — Anthropic releases Claude Fable 5 and Claude Mythos 5 to the public. [confirmed — Anthropic]

What Is Confirmed

By Anthropic's public communications:

Claude Mythos is real and publicly available as of June 9, 2026 (Fable 5 and Mythos 5).
Claude Mythos Preview is a restricted research model Anthropic considered too capable for public release, used to find zero-day vulnerabilities across major operating systems and browsers.
Project Glasswing and its launch partners — AWS, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks — plus 40+ further organizations, backed by $100M in usage credits.
Fable 5 routes high-risk queries to Claude Opus 4.8; Mythos 5 lifts those safeguards in some domains for trusted partners.

What Is Not Confirmed

Reported or leaked, not confirmed by Anthropic:

The exact Claude Mythos Preview launch date (often cited as April 7, 2026), its model ID, knowledge cutoff, and pricing ($25 / $125 reported). [reported]
Specific leaked benchmark figures and architecture or parameter estimates — see 📝Claude Mythos Benchmarks and 📝Claude Mythos Context Window. [rumor]

FAQ

Is Claude Mythos released?

Yes. As of June 9, 2026 it is publicly available as Claude Fable 5 (general availability) and Claude Mythos 5 (restricted access). A restricted research model, Claude Mythos Preview, preceded them via Project Glasswing.

What's the difference between Fable 5 and Mythos 5?

Same underlying model. Fable 5 automatically routes high-risk queries (cybersecurity, biology, distillation) to Claude Opus 4.8; Mythos 5 lifts those safeguards in some domains for vetted partners.

How much does it cost?

Fable 5 and Mythos 5 are $10 per million input tokens and $50 per million output tokens. Claude Mythos Preview pricing was higher and is not officially confirmed.

Is "Mythos" related to the MythOS platform?

No — they share only a name. See 📝MythOS vs Claude Mythos.

📝Claude Mythos — canonical reference
📝Claude Mythos for Developers — practical implications
📝Claude Mythos and AI Safety — capability-risk context

Subjective

We track Claude releases the way a carpenter tracks new tool catalogs. Mythos-class capability — now available as Fable 5 and Mythos 5 — shifts what a small team can build autonomously, which is exactly why Anthropic shipped the public model behind safety routing. We are building to take advantage of it.