OpenAI's advanced reasoning model, o3-mini, will be free for all
What's the story
OpenAI's CEO Sam Altman has announced that the company's latest reasoning model, o3-mini, will be available to all users for free.
The decision comes after the new model successfully passed external safety testing.
Both an API and ChatGPT version are expected to launch in the coming weeks.
In a post on X, Altman said, "the free tier of chatgpt is going to get o3-mini! (and the plus tier will get tons of o3-mini usage)."
Model features
o3-mini: A breakthrough in AI technology
The o3 model was first unveiled during '12 Days of OpenAI' event in December.
Greg Brockman, the company's president and co-founder, called it "a breakthrough, with a step function improvement on our hardest benchmarks."
Unlike traditional generative models such as GPT-4o Turbo, the o3 and its predecessor o1 conduct internal fact and reasoning checks before responding to user queries.
This unique capability enables them to tackle complex math, science, and coding tasks that conventional models might struggle with.
Model efficiency
o3-mini: A compact and efficient AI model
The o3-mini is a compact version of the full o3 model, which is smaller than GPT-4o Turbo.
Despite the size, it is expected to deliver faster response times and require less computational power than current models.
This efficiency makes it an ideal choice for powering ChatGPT, especially for users on the free tier.
Strategy shift
OpenAI's departure from tradition with o3-mini
Traditionally, OpenAI has powered ChatGPT with a previous-generation model while restricting access to its latest models.
However, the decision to give free users access to o3-mini marks a major departure from this strategy.
Altman said paid tier users will continue to have GPT-4o Turbo but will also see "tons of o3-mini usage."
The company has not disclosed the exact timeline for o3-mini's rollout to all users.