Openai is making a smaller, more efficient version of its smartest artificial intelligence model available for free as it seeks to answer the hype and enthusiasm around a new open source offer of the Chinese AI -Start Deepseek answered.
Wired previously reported that Openai has prepared the new model, called O3 mini, for exemption on January 31.
O3-Mini, which mocked Openai in December, is a smaller version of the model that contains the most advanced AI reasoning capabilities of any open-end offering so far. The model can divide difficult problems into constituent parts to find out how to solve it best.
“This powerful and fast model promotes the boundaries of what small models can reach,” the company said in a blog post announcing the availability of O3 mini.
Openai is making O3 mini available for all plus, team and pro-users of Chatgpt. Users of the free version of Chatgpt will also be able to try O3 mini, but will not be able to send as many questions, the company says.
Openai apparently uses PhD students to help a new model train a new model. A few weeks ago, the company began recruiting PhD computer science students at $ 100 per hour for a ‘research collaboration’ who would “work on unsaid models”, according to “Ne post seen by Wired.
Openai also appears to have recruited PhD students with expertise in other areas by a company called Mercor that he regularly uses to find staff for model training. In a recent Mercor postal placement on LinkedIn, it is said: ‘The general purpose of this project you can be part of is to create challenging scientific coding questions designed to test the capabilities of major language models to generate code to realistic scientific solving research problems. ‘
The postal placement further gives an example problem that is strikingly similar to a problem in a benchmark called Scicode designed to test a major language model’s ability to solve complicated scientific problems.
The news came when Deepseeek’s R1 still made the US technology industry disappear. The fact that such a powerful model can be released free of charge puts pressure on Google and anthropically to lower their prices.
Openai is especially eager to demonstrate that it remains at the forefront of the development and commercialization of AI, according to sources in the company.
DeepSeek’s free available model contains innovations that made it more efficient to train and serve both. It seems that the company has developed it using far fewer resources than Openai and other US businesses currently building border AI models, although the exact details of Deepseek’s expenses remain unknown. Openai says he believes that R1 may have included the output of his models in his education.
Do you have a tip?
Are you a current or former employee at Openai? We would like to hear from you. Using a non -work telephone or computer, contact will at [email protected] or on signal via its username wak01.
Openai’s latest model may not be the price of R1 in terms of price, but it shows that the company will be effectiveness of its focus in the future. Openai also says that the model is particularly strong in math, science and coding.
The company says that the latest model will also include new features, including the ability to use on web searches, link functions of a user’s code, and to link between different reasoning levels that trade the velocity of problem -solving capabilities.
The sudden rise of Deepseek also raised questions about the US government’s strategy to combat China’s rise in AI. The past two US administrations have set up a number of sanctions to combat China’s ability to access the most advanced Nvidia chips, which is usually used to build the latest AI models. Deepsheek described different types of Nvidia chips in his research, but it is unclear what exactly was used.