OpenAI is reportedly piloting a high-priced subscription tier for ChatGPT that would cost $500 each month. The plan, dubbed Pro Max, appears aimed at users who need accelerated performance from the company’s AI agents rather than additional usage limits.

Subscription tier details

Leaks from code repositories and screenshots shared by several independent observers indicate the new tier would sit above the existing $20 Plus plan and the $200 Pro tier launched in late 2024. The primary benefit listed is a speed boost for two of OpenAI’s agent services: the Work assistant, which handles multi-step research and writing tasks, and Codex, a system that generates and debugs code over extended sessions. Additional perks mentioned in the leaked material include larger memory allocation and 100 GB of file storage.

Hardware partnership

The faster response times are said to rely on a partnership with Cerebras, a chipmaker that produces wafer-scale processors the size of a dinner plate. Cerebras markets its chips as considerably faster than Nvidia’s offerings while being cheaper per unit of performance. The use of this hardware is presented as the reason the Pro Max tier can deliver a notable inference speed advantage, even though the price premium over the standard Pro plan is significant.

Market context

If confirmed, the $500 price point would place OpenAI’s top tier well above competing offerings. Anthropic’s Claude Max plan currently ranges between $100 and $200 per month, while Google’s AI Ultra plan was reduced to $100 per month earlier this year. OpenAI is expected to address the new tier at its upcoming DevDay conference on September 29, an event where the company has historically announced pricing changes.

Why it matters

The introduction of a premium, speed-focused subscription could reshape the pricing landscape for advanced AI services, signaling that providers are willing to monetize performance gains tied to cutting-edge hardware. For enterprises and developers that rely on rapid, large-scale inference for research or software development, the Pro Max tier may become a valuable, albeit costly, option. At the same time, the move could intensify competition among AI firms as they balance accessibility with the costs of maintaining high-performance infrastructure.