AI API Billing: Token-Based Pricing for LLM Products
Blog post from Lago
In June 2025, Cursor's shift from a flat-rate to a credit pool billing model for its Pro plan led to unexpected high costs for some users, prompting a public apology from the CEO and refunds for affected users. The incident highlighted the complexities of pricing products built on dynamic token-based models like GPT-4o and Claude Opus, where input and output costs can vary significantly across different models and pricing structures. The narrative explores various pricing strategies, such as passthrough, flat per-request pricing, blended pricing, and credits, each with its own advantages and challenges. It emphasizes the importance of building a robust metering and rating infrastructure to track and manage these variations, ensuring that pricing decisions are informed by accurate and timely data to prevent unexpected margin discrepancies. The text suggests that Lago's usage-based billing engine can provide the necessary visibility and flexibility for businesses to adapt their pricing strategies effectively in a fluctuating market.
No tracked trend matches for this post yet.
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.