What ChatGPT Pro’s $200 price changes for power users

OpenAI has launched ChatGPT Pro, a $200-per-month plan built around heavier access to its models, including o1 pro mode. The new tier is aimed at users who push ChatGPT on math, programming, writing, data science and other difficult tasks.

WTF Index NEUTRAL
◄ Terminator 1 Idiocracy 1 ►

This is mainly a product and pricing update for heavier AI access, with only mild implications for greater dependence on AI tools.

What ChatGPT Pro’s $200 price changes for power users

OpenAI has introduced ChatGPT Pro, a new subscription tier for ChatGPT that costs $200 per month and gives paying users a higher-end version of the company’s AI chatbot experience.

The plan centers on broad access to OpenAI’s models, especially its o1 reasoning model and a more compute-intensive option called o1 pro mode. The company is positioning the tier for people who already use ChatGPT heavily and need stronger performance on difficult work.

A $200 tier for heavy ChatGPT users

ChatGPT Pro is OpenAI’s most expensive ChatGPT plan so far. At $200 per month, it costs 10 x the cost of ChatGPT Plus, making it a product aimed at a narrow group rather than casual users.

During a livestreamed press conference on Thursday, Jason Wei, a member of OpenAI’s technical staff, described the intended audience clearly: “We think the audience for ChatGPT Pro will be the power users of ChatGPT — those who are already pushing the models to the limits of their capabilities on tasks like math, programming, and writing,”

The package includes unlimited access to all of OpenAI’s models, including the full version of o1. It also includes unlimited access to GPT-4o and Advanced Voice Mode, the conversational voice feature in ChatGPT.

That mix makes ChatGPT Pro less about one single model and more about removing limits for people who use several ChatGPT capabilities throughout their work. The tradeoff is price: many users already view ChatGPT Plus as expensive, and ChatGPT Pro raises the premium bar much higher.

Why o1 matters

The key technical change behind the plan is o1, OpenAI’s reasoning model. Unlike most AI models, o1 and similar reasoning models try to check their own work while solving a problem. That process can help avoid mistakes, but it can also make answers take longer.

OpenAI released a preview of o1 in September. The newer version is described by an OpenAI spokesperson as “a faster, more powerful, and accurate reasoning model that is even better at coding and math,” compared with the preview.

The full o1 model also adds image-upload reasoning, which was not available during the preview. OpenAI has trained it to be “more concise in its thinking” in order to improve response times.

According to OpenAI’s internal testing, o1 reduces “major errors” on “difficult real-world questions” by 34% compared to the preview version. At the same time, the source article notes an important complication: the full o1 performs worse than the preview version on several common benchmarks, including MLE-Bench, which evaluates AI agents on machine learning engineering tasks.

What o1 pro mode adds

ChatGPT Pro subscribers get access to o1 pro mode, which OpenAI says “uses more compute for the best answers to the hardest questions.” Users can select “o1 pro mode” in the model picker and ask a question directly.

Because o1 pro mode can take longer to answer, ChatGPT displays a progress bar. If a user switches away to another conversation, ChatGPT can send an in-app notification when the response is ready.

The feature appears designed for situations where speed is less important than depth and reliability. The source article notes that o1 pro mode may increase the amount of reasoning time before the model responds, and connects that to OpenAI’s earlier comments about experimenting with o1 models that reason for hours, days, or even weeks.

OpenAI says external expert testers found o1 pro mode more reliable in demanding areas. An OpenAI spokesperson told TechCrunch that it produces “more reliably accurate and comprehensive responses, especially in areas like data science, programming, and case law analysis,” and that it performs better than o1 and o1-preview on challenging machine learning benchmarks across math, science, and coding.

The company also said it saw a 75% reduction in errors for easier coding competition questions that are more reflective of everyday programming queries.

Who gets o1 without ChatGPT Pro

ChatGPT Pro is not required to use o1. As of this afternoon, all paid ChatGPT users can access o1 through the ChatGPT model selector tool.

That distinction matters. The $200 plan is not simply a ticket to o1; it is a way to access a higher-compute version through o1 pro mode, plus unlimited access to other OpenAI models and features.

For many users, the regular paid access to o1 may be enough. For users working on harder coding, math, data science, writing or legal-analysis tasks, OpenAI is betting that the extra reliability and access in ChatGPT Pro will justify the cost.

API plans, grants and the business pressure behind the move

OpenAI also plans to bring o1 to its API. The company says the API version will include new capabilities such as function calling, which allows the model to use outside tools, and image analysis. OpenAI also plans to add support for web browsing, file uploads and more in the months ahead.

The company is also giving away some ChatGPT Pro subscriptions. OpenAI announced a program to award 10 grants of ChatGPT Pro to medical researchers at “leading institutions,” with plans for more grants across “various disciplines” in the future.

The pricing also fits a broader financial story. The source article says price increases for premium ChatGPT access have long been rumored, including reporting that OpenAI expects to charge $44 per month for ChatGPT Plus by 2029. It also notes that the company has considered ultra-costly business subscriptions with more functionality and access to models under development.

OpenAI’s revenue and cost pressures are significant. The source article reports that monthly revenue reached $300 million in August, while OpenAI expects to lose roughly $5 billion this year. Staffing, office rent and AI training infrastructure are named as major expenses, and ChatGPT alone was reportedly costing OpenAI $700,000 per day at one point.

ChatGPT remains one of OpenAI’s largest revenue sources. The platform has over 300 million weekly active users, including around 10 million paying subscribers. ChatGPT Pro is therefore both a product for the most demanding users and a test of how much the market will pay for more advanced AI access.