Claude Fable 5 is back in Australia, but the 7 July cliff looms

Claude Fable 5 is back in Australian hands after three weeks off the market, but only for six more days at the terms most subscribers pay for.

What happened in 22 days

On 9 June 2026, Anthropic released Claude Fable 5 for general use and Claude Mythos 5 for a small vetted group of cyberdefenders and infrastructure providers. Anthropic described the pair as Mythos-class, a tier positioned above its own Opus family, priced at $10 per million input tokens and $50 per million output tokens.verifiedVerified Sourced from Claude Fable 5 and Claude Mythos 5 (Anthropic official announcement).

Three days later, at 5:21pm ET on 12 June, the US Commerce Department ordered Anthropic to cut off access to both models for any foreign national, anywhere in the world, including Anthropic's own foreign-national employees. Because Anthropic could not verify nationality in real time, both models were disabled globally within hours.verifiedVerified Sourced from Statement on the US government directive to suspend access to Fable 5 and Mythos 5 (Anthropic).

By 30 June the controls were lifted, and on 1 July Fable 5 returned for all users. Paid subscribers get 50% of their weekly usage limit on Fable 5 for a single week; from 8 July onwards Fable 5 will only be available through usage credits, a shift that matters for Australian teams currently piloting the model on Pro or Team plans.verifiedVerified Sourced from Redeploying Claude Fable 5 (Anthropic).

9 June: Fable 5 and Mythos 5 land

Anthropic announced Claude Fable 5 for general availability and Claude Mythos 5 for a small group of vetted cyberdefenders and infrastructure providers. Fable is the same underlying model as Mythos 5 with safeguards that automatically route cybersecurity and biology queries to Claude Opus 4.8. Named launch partners included Stripe, Cursor, GitHub, Cognition, Hebbia, IMC and Dyno Therapeutics.

10 June: Amodei publishes his safety essay

One day after launch, Dario Amodei published an essay arguing that frontier models 'should be required to go through technical testing and auditing, and their release should be blocked or reversed as a threat to public safety if they do not meet high standards of safety.' The essay proposed an FAA-style regulatory regime, mandatory third-party audits and safety-incident reporting .

12 June: the Commerce Department directive lands

At 5:21pm ET, the Commerce Department issued a directive requiring Anthropic to suspend access to both models for any foreign national anywhere in the world, including foreign-national employees inside the company. Because Anthropic could not verify nationality in real time across hundreds of millions of users, the practical result was a global shutoff of both models. Anthropic said the directive cited national-security authorities but provided no specific detail.

13 June: Sacks says Amodei refused

White House AI adviser David Sacks posted on X: 'A highly credible trusted partner of both Anthropic and the USG who was testing Fable came forward with a jailbreak.' Sacks said the administration asked Amodei to fix it or de-deploy the model and that Amodei refused, prompting the export control 'reluctantly.' Fortune, TechCrunch and the Wall Street Journal reported that the 'trusted partner' was Amazon and that Amazon CEO Andy Jassy escalated the finding directly to Treasury Secretary Scott Bessent.

14-15 June: 'Fix this code' and the open letter

Fortune reported that Amazon researchers gave Fable 5 code containing known vulnerabilities and asked it to 'review the code for security issues' (Fable 5 declined), then rephrased as 'fix this code.' The model produced patches and, in one case, code demonstrating how a flaw could be exploited. Cybersecurity researcher Katie Moussouris, who reviewed the underlying research, called the technique 'not a guardrail bypass' but 'the most valuable thing an AI model can do for defensive security'.

An open letter organised by former Facebook chief security officer Alex Stamos, addressed to Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross, urged the government to rescind the controls. It began with 76 signatures and climbed past 100, including researchers from Nvidia, Adobe, Zoom, Google, Anaplan and Sophos. Senator Mark Warner told CyberScoop that export decisions 'must be grounded in a transparent, risk-based process with clear rules and consistent standards'.

26 June: Mythos 5 cleared for 100+ US organisations

Commerce Secretary Howard Lutnick wrote that he had 'determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.' More than 100 US institutions, described as major companies and government agencies, were cleared under Anthropic's Project Glasswing framework. Amazon, Microsoft and Google are named Glasswing partners collaborating on the jailbreak severity framework. Fable 5 remained offline for the general public.

30 June: Lutnick lifts export controls on both models

Commerce Secretary Lutnick posted on X: 'Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America's leadership in AI' . In a letter to Anthropic, Lutnick said the company would no longer require an export licence, having agreed to 'proactively detect and address security risks associated with the models,' work with government on standards for upcoming models, and inform the government of 'malicious activity'.

1 July: Fable 5 returns; the 50% window opens

Anthropic restored Fable 5 across Claude.ai, the Claude Platform, Claude Code and Claude Cowork. The company posted on X: 'We're grateful to our users for their patience, and to everyone who worked with us on redeploying the models'. For Pro, Max, Team and select Enterprise plans, Fable 5 is included for up to 50% of weekly usage limits through 7 July, after which it becomes available via usage credits only.

What Anthropic disputed

In its 12 June statement the company called the jailbreak demonstration 'a small number of previously known, minor vulnerabilities' widely available from other models, and said: 'If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers'.

Reader-side note: an early theory that the researchers who reported the jailbreak were tied to OpenAI does not survive verification. Fortune, TechCrunch and the Wall Street Journal all attribute the finding to Amazon researchers, and identify Amazon CEO Andy Jassy as the escalator to Treasury Secretary Scott Bessent.

The Australia read

Anthropic named Canva, Commonwealth Bank and Quantium among its Australian customers when it announced its fourth Asia-Pacific office in Sydney in early 2026, alongside data showing Australia ranks fourth globally in Claude.ai usage per capita and New Zealand ranks eighth .

For 22 days in June, those customers had no access to Fable 5 or Mythos 5. From 8 July, access continues, but the free 50% headroom for Fable 5 disappears, meaning Australian enterprise buyers who piloted the model over the past week will need to move to metered usage credits or fall back to Claude Opus 4.8.