AI Frontier Shifts: Fable 5 Returns, Square’s AI Ordering, and Global Impacts

The AI frontier continues to move at speed, mixing breakthroughs with sharp regulatory turns. After a tense few weeks, Anthropic’s Claude Fable 5 and Mythos 5 are back online, though with different accessibility paths. U.S. authorities lifted the emergency export controls on Fable 5, restoring global access, while Mythos 5 remains behind a vetted-access regime awaiting broader government approvals. Enterprises are already weighing how to deploy multiple models, govern data, and ensure safety as they modernize workflows.

Behind the scenes, pricing and safeguards are shaping adoption. Anthropic has priced Fable 5 and Mythos 5 at $10.00 per million input tokens and $50.00 per million output tokens, the costliest frontier options on the market. There’s also a 30-day data-retention requirement, which firms must weigh against privacy mandates. At the same time, the company is offering temporary promotions through July 7 that bundle Fable 5 usage with certain plans, then transition users to credits thereafter.

Meanwhile, a broader ecosystem is pushing forward with new approaches to delivery and code generation. Square’s new agentic-commerce integration lets restaurants accept orders placed directly via ChatGPT or Claude screens, avoiding marketplace fees and letting orders flow straight into existing Square workflows. That experiment promises to shift economics in food service, especially when contrasted with typical 25–30% commissions charged by major delivery platforms.

In imaging and vision, Google has expanded its AI imaging offerings with Nano Banana Lite 2 alongside Gemini Omni Flash, highlighting how faster, higher-quality visuals are becoming a standard feature of frontier platforms. These tools complement the earlier LLM breakthroughs by enabling richer media processing for business and creative teams alike.

Yet the AI revolution is not just a technology story; it’s a global policy conversation. A United Nations report warns that rapid AI adoption could worsen global inequality unless a shared framework for responsible development is adopted. The push toward sovereignty—more local and open-weight models, hardware- and governance-led approaches—contrasts with the U.S.-centric, export-control-driven path of recent months. This tension is echoing in board rooms and regulatory briefings as enterprises plan for resilience through model-agnostic architectures and fallback strategies.

As the landscape evolves, observers note that the AI race is now a multi-faceted negotiation among nations and companies. From Commonwealth prize verdicts to the ongoing chatter about copyright and datacentre funding, the industry is recalibrating what it means to deploy frontier tech in production. The practical takeaway for readers is clear: diversify models, maintain governance, and prepare for both rapid deployments and careful compliance. In an era where a single model facing export controls can ripple across supply chains, resilience means having alternatives at the ready—including local, open-weight options alongside the leading closed-platform models.

1. Anthropic brings back Claude Fable 5 globally after US export controls lift
2. We can live without AI, but can we live without clean water? | Guardian Letters
3. Restaurants can now accept orders placed directly from ChatGPT and Claude thanks to Square’s new integration
4. Google Expands AI Imaging Offerings With Nano Banana Lite 2
5. Short story accused of being AI-written wins Commonwealth prize
6. Anthropic says US has lifted export controls on Fable and Mythos AI models after security fears
7. Rapid spread of AI may worsen global inequality, UN warns
8. Creatives sound alarm on copyright as Pocock calls $50bn datacentre proposal ‘ultimate dirty deal’