The new Openai GPT 4.1 models differ in encoding

OPENAI announced today that he is launching a new family of artificial intelligence Models optimized to distinguish themselves in coding as it is intensifying its efforts to repel more firm competition from companies such as Google and Anthropic. The models are accessible to developers through the OPENAI (API) programming interface.

Openai launches three sizes of models: GPT 4.1, GPT 4.1 Mini and GPT 4.1 Nano. Kevin Weil, Chief Product Officer at Openai, told Livestream that the new models are better than the most used model of Openai, GPT-4O and better than its largest and most powerful model, GPT-4.5, somehow.

The GPT-4.1 marks 55 percent of the SWE-Tala, a widely used indicator of measuring the skill of encoding models. The result is several percentage points over those of other Openai models. The new models are “great in encoding, they are great in sophisticated instructions after, they are fantastic for construction agents,” said Weil.

The capacity of AI models to write and edit code has improved significantly in recent months, allowing more automated ways to prototyping the software and improve the capabilities of so-called AI agentsS Rivals as Anthrop and Google Both introduced the model that are particularly good at writing code.

The arrival of the GPT-4.1 has grown for weeks of weeks. Openai has apparently tested the model of some popular charts under the alphan alpha Quasar, sources say. Some users of the stealth model reported Impressive encoding abilities. “Quasar fixed all the open problems I had with other codes generated (SIC) through LLMS that were incomplete,” writes a Reddit man.

All new models can analyze eight times more code at a time, which improves their ability to make improvements and correct errors. The new models are also better with the following instructions given by users, reducing the need to repeat commands in different ways to achieve the desired result. Openai showed GPT-4.1 demonstrations, building various applications, including flash language learning cards application.

“Developers take a lot of care of encoding and we are improving our model’s ability to write a functional code,” said Michelle Parakras, who works after Openai training during Livestream on Monday. “We work to follow it different formats and better examine repositions, start tests on units and write a code that compiles.”

GPT-4.1 is 40 percent faster than GPT.4O, the most widely used Openai model for developers. The price of users who enter requests are reduced by 80 percent in this latest version, says Openai.

On today’s Livestream Varun Mohan, CEO of Windsurf, a popular AI encoding tool, said the company tested GPT-4.1 and found that the new model was “60 percent” better than GPT-4O according to its own indicators. “We have found that GPT-4.1 has significantly fewer cases of degenerated behavior,” Mohan said, noting that the new model spends less time to read and edit non-license files by mistake.

In the last few years OpenAI has applied a feverish interest in Chatgptchatbot First discovered at the end of 2022.In a growing business selling access to more modern chatbots and AI models. In an interview with TED last week, Altman said Openai has 500 million active users a week and that the use is “growing very fast”.