Anthropic announced Wednesday the launch of a premium subscription tier, Claude: Max, priced at $200 per month. The new offering is positioned as a direct competitor to OpenAI’s similarly expensive ChatGPT Pro subscription, both targeting AI power users who demand extensive access and advanced features.
Claude: Max notably raises usage limits far beyond the existing Claude Pro plan, which is priced at $20 per month. Interestingly, the “Max” subscription is available in two different tiers: a $100 monthly package providing usage levels five times higher than Claude Pro, and the top-tier $200 option featuring a twenty-fold increase over the base subscription.
Leading AI model developers are increasingly turning to higher-priced subscription models as a strategy for significant revenue growth. Following OpenAI’s introduction of ChatGPT Pro, the company reportedly experienced a major uptick in annualized revenue, suggesting there is substantial market opportunity for such premium tiers. Anthropic is clearly aiming for a similar trajectory with its new Max subscription.
Scott White, Anthropic’s product lead, noted in an interview that the firm has not ruled out adding even more expensive subscription packages in the future. Unlike OpenAI, Anthropic currently does not offer an unlimited usage subscription option. When asked about the possibility of launching a $500-per-month Claude plan, White confirmed that Anthropic is open-minded and will continue exploring various options based on user demand and feedback.
Given the steep ongoing costs of developing advanced AI models, Anthropic is actively diversifying its revenue streams. Recently, the company introduced “Claude for Education,” a specialized subscription aimed at universities and institutions, which provides unique tools tailored to educational environments.
While the company declined to disclose exact subscription numbers, White said the demand for the latest AI model, Claude 3.7 Sonnet, has been significant. Claude 3.7 Sonnet is Anthropic’s first specialized reasoning model, designed to deliver more reliable responses by leveraging significantly more computing power than their traditional models.
