Microsoft Slashes AI Costs with Token Budgets and Default Model Switch

Microsoft is implementing new limits on its employees' use of artificial intelligence (AI) due to skyrocketing costs. According to an internal email obtained by 404 Media, the company is making OpenAI GPT-5.6 its default model for internal use to 'get greater value from our token investment.' This change aims to reduce excessive spending on AI tokens, which are like credits that allow employees to prompt AIs with specific tasks.

In a similar move, other companies such as Amazon, Adobe, Atlassian, and Citi have also cracked down on their employees' token spending. Microsoft's EVP Jay Parikh emphasized the need for employees to be aware of how they consume tokens, stating that 'tokenmaxxing' is not what the company aims to optimize for.

Instead, Microsoft wants its employees to focus on maximizing outcomes that benefit customers and the business. To achieve this goal, divisions will have an AI token budget target, which employees can track individually. This new approach is part of the company's effort to manage token spend with 'the same discipline we apply to every other critical resource.'

The move comes after GitHub Copilot moved to usage-based billing two months ago, causing some users to quickly hit their limits. Microsoft is also attempting to cut down on AI costs by sending Microsoft 365 AI prompts to its internal MAI models rather than relying on Anthropic and OpenAI.