Major AI Providers Shift to Consumption-Based Token Pricing
Anthropic, OpenAI, Google, and GitHub replaced flat-fee monthly pricing with token-metered billing to address the economic unsustainability of flat-rate AI services.
Major AI providers, including Anthropic, OpenAI, Google, and GitHub, dismantled flat-fee monthly pricing structures in 2026. These companies replaced predictable software-as-a-service expenses with consumption-based, token-metered billing, effectively treating AI tooling as a variable cost similar to cloud infrastructure.
The industry-wide transition follows a period where flat rates were used primarily for customer acquisition but proved economically unsustainable for providers. As a result, OpenAI transitioned its federal OneGov program to token-metered pricing, while Google shifted Gemini paid plans from daily prompt caps to compute-based limits. GitHub similarly moved GitHub Copilot toward consumption-based pricing.
This shift has introduced significant financial volatility for enterprise users. A 2026 survey by Mavvrik found that 62% of organizations experienced unexpected AI cost spikes, driven primarily by usage overages. To manage these costs, engineering leaders are adopting hybrid strategies. They now route high-volume, routine tasks to cost-effective open-weight models, such as DeepSeek V4 Pro and Alibaba's Qwen3-Thinking-2507, while reserving expensive closed frontier models for complex reasoning tasks.
