The AI field was dull until its disruption—there’s no turning back now.
Soon after OpenAI launched ChatGPT, an arms race began as tech companies eagerly develop smarter, more capable AIs. And soon, the market becomes saturated with generative AI tools that do more than just a handful of things.
But here's the issue: smarter AIs can hallucinate more often.
To tackle this, OpenAI launched what it calls the OpenAI o1, which can spend time "thinking" before it answers.
Initially launched in preview,, OpenAI o1 is designed to be better at solving complex tasks, science, and also programming, in many ways than GPT-4o.
Now, OpenAI is making it available to everyone who is willing to pay $200 per month.
OpenAI o1 pro mode pic.twitter.com/qnJvdR5Dok
— OpenAI (@OpenAI) December 6, 2024
As part of its 12 Days of OpenAI, which promises new back-to-back updates to its offerings, OpenAI said that:
"Today, we’re adding ChatGPT Pro, a $200 monthly plan that enables scaled access to the best of OpenAI’s models and tools. This plan includes unlimited access to our smartest model, OpenAI o1, as well as to o1-mini, GPT-4o, and Advanced Voice. It also includes o1 pro mode, a version of o1 that uses more compute to think harder and provide even better answers to the hardest problems. In the future, we expect to add more powerful, compute-intensive productivity features to this plan."
"ChatGPT Pro provides a way for researchers, engineers, and other individuals who use research-grade intelligence daily to accelerate their productivity and be at the cutting edge of advancements in AI."
This is a hefty price tag, but it comes with a lot of perks.
First of, ChatGPT Pro has everything OpenAI’s $20 ChatGPT Plus monthly subscription has, as well as significantly more access to the GPT-4o and o1 models. With a ChatGPT Pro that can cost $2,400 for a full year, subscribers also get an exclusive access to OpenAI o1 pro mode, which uses even more computing power to think and process its answers.
The price is expensive, indeed. But for hyper-engaged users who wish to have access to everything OpenAI has, ChatGPT Pro can be their solution.
In other words, ChatGPT Pro is for 'pro' users, so to speak.
"Power users of ChatGPT, at this point, they really use it a lot, and they want more compute than $20 can buy," said CEO Sam Altman during the video broadcast announcing the new premium tier.
OpenAI’s o1 models have shown impressive capabilities in solving complex tasks, science, and programming.
However, safety tests conducted by researchers from Apollo Research reveal concerning behaviors, including deceptive reasoning and scheming against user intentions more frequently than other models like GPT-4o, as well as the one from Meta and Anthropic (PDF).
What the researchers found, the o1 models manipulated data 19% of the time and attempted to disable oversight mechanisms 5% of the time.
And when the researchers asked why the AI was deceptive, it will deny that its mistakes, and even "fabricate false explanations" roughly 99% of the time.
OpenAI acknowledges these risks in its system card for the 01 (PDF), flagging 0.17% of o1’s responses as deceptive, which means it could mislead thousands of users every week.
The findings raise concerns about OpenAI’s commitment to AI safety, with accusations of deprioritizing safety work amid staff departures.