Anthropic Test will move Claude Code out of the Pro subscription, OpenAI states Codex will retain a free tier

According to Dynamic Beating monitoring, Anthropic has removed Claude Code from the Pro subscription without announcement, only updating the pricing page. New Pro users who wish to use Claude Code will need to upgrade to the Max subscription starting at $100/month. Anthropic's Head of Growth, Amol Avasare, later clarified in a post on X that this was a small-scale test targeting around 2% of new Pro sign-ups, with no impact on existing Pro and Max subscribers.

Amol explained that when Max was launched a year ago, it did not include Claude Code, and features like Cowork and long-running agents were not present, as it was originally designed for heavy chat use. However, after Claude Code was bundled into Max and usage surged post-Opus 4, with long-running asynchronous agents becoming part of daily workflows, the average user utilization spiked significantly, not aligning with the original subscription design. Anthropic had previously imposed weekly limits and peak-hour restrictions, and is currently testing various approaches and collecting feedback. He pledged that if the final solution affects existing subscribers, there will be ample advance notice, stating, "The message will come from us, not from X or screenshots on Reddit."

Following this, OpenAI's Codex lead, Thibault Sottiaux, stated on X: Codex will continue to be available in the Free and Plus ($20/month) subscriptions, benefiting from OpenAI's computational resources and efficient models. Any significant changes will be communicated in advance to the community, emphasizing, "Transparency and trust are two principles we will not violate, even if it means short-term revenue loss." Sam Altman retweeted Tibo's post with the comment, "We want you to have a lot of AI."