Openai has announced today that he releases a modern family of models of artificial intelligence optimized to lead in coding, because it increases efforts to repel more and more unyielding competition from companies such as Google and Anthropic. Models are available to programmers via the OPENAI (API) application programming interface.
Openai releases three sizes of models: GPT 4.1, GPT 4.1 Mini and GPT 4.1 Nano. Kevin Weil, product director at Opeli, said on live broadcast that modern models are better than the most commonly used OPENAI, GPT-4O and better than its largest and most powerful model, GPT-4.5, in some respects.
GPT-4.1 obtained 55 percent on Benche of its, commonly used reference point for assessing the efficiency of coding models. The result is several percentage points above the result of other OPENAI models. The modern models are “great in coding, they are great in complex instructions, they are fantastic for building agents,” said Weil.
The ability of AI models to write and edit the code has improved significantly in recent months, enabling more automated methods of prototyping software and improving the skills of so -called AI agents. Rivals such as Antropic and Google have introduced models that are particularly good in writing the code.
The arrival of GPT-4.1 has been widely rewarded for weeks. Opeli apparently tested the model on some popular leaders’ boards under the nickname Alpha Quasar, say sources. Some users of the “hiding” model Reported Impressive coding capacity. “Quasar repaired all open problems that I had with other genres code [sic] Through LLMS, which was incomplete, “wrote one person at Reddit.
All modern models can analyze eight times more code at the same time, which improves their ability to introduce improvements and repair errors. The modern models are also better to follow the instructions given by users, reducing the need to repeat commands in various ways to get the desired result. Opeli showed a GPT-4.1 demo building of various applications, including the Flashcard application for language learning.
“Developers take care of coding very much, and we improve the ability of our model to write a functional code,” said Michelle Pokrass, who is working on training at Opeli, live on Monday. “We worked on tracking various formats and better explore repo, run unit tests and write the code that compiles.”
GPT-4.1 is 40 percent faster than GPT.4O, the most commonly used OPENAI model for programmers. Opeli says that the cost of entering the query has been reduced by 80 percent in this latest version.
In today’s live broadcast Varun Mohan, CEO Windsurf, a popular AI coding tool, said that the company tested GPT-4.1 and stated that the modern model was “60 percent” better than GPT-4O in accordance with its own comparative tests. “We found that GPT-4.1 has much less cases of degenerated behavior,” Mohan said, noting that the modern model spends less time reading and editing irrelevant files by mistake.
Over the past few years, Opennai Parlaya, the feverish interest in Chatgpt, an unusual chatbot for the first time presented at the end of 2022, in the developing business selling access to more advanced chatbots and AI models. In an interview with TEDs last week, Altman said that Opeli had 500 million weekly users and that the operate of “grew very fast.”