OpenAI launches $200 ChatGPT Pro plan with unlimited o1

OpenAI just dropped its most expensive subscription tier yet, asking power users to pay $200 every month for unlimited intelligence.

Rolled out on December 5, ChatGPT Pro costs $200 a month in the US, £200 in the UK, and €229 in Europe. Current Plus subscribers who decide to make the jump will get prorated credits for their remaining days.

For ten times the price of Plus, OpenAI is giving subscribers unlimited access to its smartest models and a raw compute boost aimed straight at solving the hardest human problems.

Giving o1 more room to think

Subscribing to Pro completely removes the usage caps. You get unlimited turns with the full release of OpenAI o1, the distilled o1-mini, GPT-4o, and Advanced Voice Mode.

Alongside that freedom comes o1 pro mode. Basically, it lets the o1 model burn significantly more compute and spend extra time reasoning through tasks before giving an answer.

External expert testers found that this mode delivers far more reliable and thorough answers in complex fields like data science, advanced programming, and case law analysis.

OpenAI proved the difference on benchmarks like AIME 2024 math, Codeforces competitions, and the doctoral-level GPQA Diamond. To demonstrate its reliability, OpenAI even introduced a brutal evaluation rule: a problem is only marked correct if the model answers it right four out of four times. In Codeforces, scores were based on the worst-performing solution of the four.

Because this deep-thinking process takes real time, ChatGPT displays a visible progress bar while running. If you switch to another chat window, it sends an in-app notification once the answer is ready.

Before the official announcement, software engineer Tibor Blaho discovered leaked pages revealing the Pro tier. Interestingly, those leaks also mentioned an upcoming limited preview of GPT-4.5 for ChatGPT Team subscribers, though OpenAI quickly took down that reference.

OpenAI is also giving away ten ChatGPT Pro grants to prominent US medical researchers studying rare diseases, aging, and cancer immunotherapy at institutions like Harvard Medical School and Berkeley Lab.

Why it matters

We think ChatGPT Pro signals a fundamental shift in how frontier AI is priced and packaged.

For everyday users and casual builders, this plan simply makes no sense right now. The $200 monthly sticker shock is steep, and unless you are an engineer writing complex infrastructure code or a researcher processing intricate scientific data, existing tiers are more than enough.

For the AI industry, OpenAI is effectively decoupling simple chat queries from heavy compute-intensive reasoning. Rather than trying to cram ever-increasing test-time compute into a subsidized $20 plan, it is creating an elite tier to see who will truly pay for PhD-level reliability.

It is a clever way to subsidize high inference bills without alienating the general public, but the stakes are much higher now. When people pay ten times the normal rate, their tolerance for hallucinations drops to zero. If o1 pro mode stumbles on a high-stakes task, justifying that $200 invoice will get very difficult very quickly.