OpenAi launches a new O3 -Mini model -This is how free chatgpt -users can try it

On the last day of OpenAi’s 12 days ‘Shipmas’, ” The company revealed its latest models, O3 and O3-Mini, which excel in reasoning and even surpass O1 In a series of benchmarks, including mathematics and science. During the launch, OpenAi CEO Sam Altman said that O3 was planned to fall at the end of January, and today the company fulfilled its promise.

O3-Mini

On Friday, OpenAi issued The O3-mini model, the most cost-efficient model in the Reasoning Series of OpenAi, for the public. So far, that series consists of O1 and O1-Mini. Just like its predecessor, the model is particularly strong in science, mathematics and coding, according to the company.

OpenAI O3-Mini is now available in Chatgpt and the API.
Pro users have unlimited access to O3-Mini and Plus & Team users have triple tariff limits (versus O1-Mini).
Free users can try O3-Mini in Chatgpt by selecting the Reason button under the composer of the message.

– OpenAI (@Openai) January 31, 2025

When O3-Mini is selected, this average reasoning effort will use, which keeps speed and accuracy in balance. Although the original O1 model still has a broader general knowledge than O3-Mini, the big advantage of the new model is the higher speed and higher performance compared to O1-Mini.

Benchmark performance

When comparing the performance of O3-Mini with O1-Mini, expert testers discovered that O3-Mini delivered more accurately, reasoned and clearer reactions than O1-Mini. According to the post, they preferred O3-mini reactions 56% of the time and saw a 39% reduction in large errors.

In addition to human preference evaluations, in various voice benchmarks, including the Math (AIME 2024), PhD students, level of scientific questions (GPQA Diamond) and competition code (CodeforCES), O3-mini with medium reasoning-what chatgpt users will be defined by default O1-Mini.

It is also remarkable that O3-Mini, with high reasoning effort in the benchmarks, came close to O1 performance, sometimes even surpass, as shown in the AIM 2024 above and software engineering (SWE-Bench verified) benchmarks. The O3-mini model with average reasoning effort corresponded to the performance of O1 in the code forces benchmark.

Safety

OpenAI rated the safety of O3-Mini through public release by Jailbreak and non-permitted content evaluations. The company discovered that the GPT-4O model considerably surpasses evaluations. OpenAi placed the evaluation results below and also launched an O3-mini system card, a 37 pages PDF This includes the detailed results of the evaluations.

How to access

All subscribers of the paid levels of OpenAi, including Chatgpt Plus” TeamAnd Procan access OpenAi O3-Mini from today. Plus and team users now have the interest limit three times, which range from 50 messages per day with O1-Mini to 150 messages per day. Chatgpt Enterprise Access comes within a week.

Also: The powerful new ‘Thinking deeper’ function of Copilot is free for all users – how it works

The O3-Mini model will replace O1-Mini in the model picker, because it would be useful for the same tasks, except that the experience will now be improved with lower latency and higher speed limits. As a paid user, I did not yet have access to the O3-Mini at the time of writing and I still see the O1-Mini option instead.

If you do not have a subscription, no worries: you can see if O3-Mini is worth the hype of your free account. The only thing that free chatgpt users have to do is click on “reason” in the message text box or regenerate an answer. OpenAI CEO Sam Altman confirmed free access to one Post on X. So far, all reasoning models have been held behind a Paywall; OpenAi has not given any restrictions around the new model for free users.