The OpenAI technological firm has launched its new O3-mini reasoning model, which is the “most profitable” which has announced to date, has an “exceptional” performance in Stem disciplines and incorporates reasoning capabilities for free users.
O3-mini offers performance comparable to Openai O1 in mathematics, coding and science, but With a “significantly lower” price and faster response times, as explained by the company in a statement.
The technology company has stressed that this new model is the first small size that admits “very requested” developer functions, including the call of functions -which allows the models to obtain data and perform actions -indications for developers or structured outputs , attached to a JSON scheme. Likewise, as OpenAi O1-mini and OpenAi O1 Preview, he will admit ‘streaming’.
This model, which is “the most profitable” because it has also reached free users, you can adjust to different levels of reasoning effort (With the options low, medium and high) to optimize specific use cases, thanks to which it can “think more” when addressing complex challenges.
On the other hand, O3-MINI does not support vision capabilities, so it urges developers to use OPENAI O1 to execute visual reasoning tasks. It also responds 24 percent faster than O1-mini, maintaining exceptional performance in Stem (science, technology, engineering and mathematics).
In comparison with this one, external evaluators prefer the responses of the new model by 56 percent more of the time and detect 39 percent less significant errors in complex questions of the real world. Likewise, the firm has advanced that O3-mini exceeds OpenAi O1 in various coding and medium level reasoning tasks.
On the other hand, the company has indicated that one of the key techniques they have used to teach the model to respond safely It is the deliberative alignment, in which they have trained the model to reason about the security specifications written by humans before responding to the user’s instructions.
Openai O3-mini, optimized for production environmentsis already being implemented in ChatgPT and in the application programming interfaces (API) of completion of conversations, attendees and lots.