OpenAI appears to be developing a new subscription tier called ChatGPT Pro Max, priced at $500 per month, according to code strings and screenshots circulating this week. Independent researcher Tibor Blaho first identified the changes in hidden logs, followed by leak tracker TestingCatalog and user mobile dev, who shared a screenshot confirming the price point. The plan is positioned as an upgrade over the existing $200 Pro tier launched in December 2024, costing 25 times more than the standard $20 Plus plan.
The primary differentiator for Pro Max seems to be speed rather than usage limits. Screenshots reviewed by outlets tracking the leak indicate the tier offers "Fastest Work and Codex," referring to OpenAI’s multi-step research agent and its autonomous coding tool. Additional reported perks include maximum memory allocation and 100GB of file storage. Some leaks suggest the service may utilize Cerebras hardware for inference, though questions remain regarding the pricing structure given Cerebras’ claims of providing 21x faster performance at one-third the cost of Nvidia B200 chips. OpenAI has not publicly confirmed the tier, but its DevDay conference is scheduled for September 29, a traditional venue for unveiling new pricing structures.
The emergence of a $500 monthly tier signals a strategic pivot toward monetizing high-performance infrastructure for professional workflows rather than general consumer chat. By isolating speed enhancements for specific agentic tools like Work and Codex, OpenAI is segmenting its market to capture value from users whose productivity depends on reduced latency. This move places OpenAI significantly above competitors, with Anthropic’s top Claude Max plan ranging from $100 to $200 and Google’s AI Ultra plan recently cut to $100, suggesting OpenAI believes it can command a premium for specialized computational resources.
However, the reliance on leaked information creates uncertainty regarding the actual value proposition. If the tier indeed leverages Cerebras hardware, which markets itself as cost-efficient relative to Nvidia alternatives, the steep price increase raises questions about margin preservation versus infrastructure costs. The lack of official confirmation ahead of DevDay leaves room for adjustment, but the consistent appearance of the $500 figure across multiple independent sources implies the pricing model is already integrated into development builds. Market observers will watch whether this tier successfully converts power users or if the price point proves prohibitive outside enterprise contexts.
