Anthropic and OpenAI announced competing pricing changes that have intensified the contest for AI-powered coding-agent users. Anthropic said it would restore support for third-party agent harnesses on paid Claude plans, but placed that usage behind a separate credit meter. Axios framed the broader development as evidence that "all-you-can-eat AI subscriptions may not survive the agent era", noting that autonomous coding agents can generate thousands of requests, run tests continuously, browse the web, and recursively call models . These workloads have no equivalent in human-driven usage patterns.