Tokenmaxxing is so over. It's all about modelmaxxing now.

In the fast-evolving landscape of artificial intelligence, a new trend is emerging: model switching. This approach allows companies to optimize their AI usage by leveraging a combination of different models based on the complexity of tasks. Morgan Linton, Chief Technology Officer at Bold Metrics, shared insights into how his team effectively utilizes various AI models, directing engineers to switch between platforms like Claude, GPT-5.5, and Cursor to achieve the best results without overspending on tokens. The term 'tokenmaxxing' dominated discussions in the AI community during the first half of 2026, as organizations encouraged employees to maximize their use of AI tools. However, a reconsideration of AI budgets has led firms, including giants like Uber and Microsoft, to adopt a more strategic stance. Professionals across various roles are discovering the advantages of directing complex tasks to premium models while relegating simpler, repetitive assignments to more cost-effective options. This shift in strategy comes in response to rising operational costs. For instance, Coinbase's CEO Brian Armstrong highlighted that within 12 to 18 months, a significant portion of workloads—approximately 80%—will shift to models that are 99% cheaper. Only a minority of tasks will require the latest, more expensive models, where precision is crucial. Many professionals have learned to adapt their workflows to maximize efficiency. Tanvi Pisal, a user-experience designer, shared her experience of initially overusing advanced models, leading to wasted tokens. Now, she adopts a design-first approach using tools like Figma to guide her use of AI, significantly optimizing her token expenditure. Moreover, companies are increasingly utilizing model routing startups to streamline the selection process. These platforms help designate tasks to the most suitable models based on their complexity, ensuring better resource management. As awareness of model routing grows, the number of firms employing these systems is on the rise, from 1% last year to 5% this year. Behavioral economics expert Dan Ariely noted that this shift reflects a broader psychological response to scarcity, similar to the old limitations on cellphone minutes. Businesses are now more cautious about their AI usage, leading to a more calculated approach regarding which models to deploy. Ultimately, as the AI landscape continues to evolve, the embrace of model switching represents a significant move towards more efficient and cost-conscious use of AI technologies. This trend is not just about leveraging the latest tools but also about understanding and optimizing the strengths of various models for specific tasks.