Australians can no longer access Anthropic's most advanced AI models after a US government directive forced the company to block foreign nationals, including Australians.
The restriction affects Fable 5 and Mythos 5, Anthropic's flagship frontier models. Mythos 5 had only been available to select Australian organisations through Project Glasswing, released earlier this month. Both models were pulled over the weekend.
Anthropic, a San Francisco AI company founded in 2021 by former OpenAI researchers, competes directly with ChatGPT and Google's Gemini. The company is backed by Amazon and Google and operates as one of the best-capitalised private AI labs globally.
The official reason: safety guardrails. The real context: advanced AI models are now treated as strategic technology, subject to export controls similar to semiconductor chips. That means access can be revoked without notice, even for allies.
For ANZ sales teams selling AI-enabled products, this matters. Anthropic has material Australian adoption: the company's own research shows Australians use Claude at rates well above what population size alone would predict. That user base just got cut off from the vendor's top-tier models mid-deployment.
Fable 5 is Anthropic's general-purpose reasoning model, designed for software engineering, research, and technical analysis. Mythos 5 is its more advanced variant, aimed at complex problem-solving with minimal human input. Both were positioned as direct competitors to OpenAI's latest releases.
Anthropic has no publicly disclosed ANZ office or headcount. The company distributes globally through cloud platforms, which means local teams relying on Claude could find themselves locked out of model upgrades without local support infrastructure.
Worth noting: this is not a vendor decision. It is a government mandate. But the outcome is the same: Australian users lost access to tools they were actively using, with no clear timeline for restoration.
For go-to-market teams evaluating AI vendors, the lesson is clear. Model access is now geopolitical risk, not just technical capability. If your stack depends on US frontier AI, build contingency.