Main to remember
- OPENAI has unveiled the GPT-4.1 series promising major coding improvements, the following instruction and the understanding of the context.
- The GPT-4.1 models offer reduced latency and improved profitability compared to previous versions.
Share this article
OPENAI has officially published its range of new generation models-GPT-4.1, as well as GPT-4.1 Mini and GPT-4.1 NANO-operating major improvements in coding, the following instruction and the understanding of the context.
The deployment follows a declaration earlier this month of the CEO of Openai, Sam Altman, revealing the plans of the company to publish the models O3 and O4-Mini “in a few weeks”.
Shortly after his remarks, the penis reported that Optai was preparing to imminently launch his new AI models. The effort is part of the wider strategy of the company aimed at improving its current offers before the planned version of the GPT-5 in 2025.
According to Openai, the new GPT-4.1 models surpass the previous versions through key references. The flagship model, GPT-4.1, scores of 54.6% on Swe-Bench verified for coding tasks-an improvement of 21.4 percentage of points compared to GPT-4O.
On the following instruction, GPT-4.1 also shows major gains. It surpasses GPT-4O of 10.5 points on the multichallenge reference and demonstrates a better understanding of the complex and multi-stage instructions, even when the prompts include formatting, tone or sequencing constraints.
Regarding context management, the three models can manage up to 1 million tokens, against 128,000 in the previous generation. OPENAI says that the models operate reliably on the complete context window, recovering and processing relevant information with greater consistency.
“We have formed these models by emphasizing public services in the real world,” said the company in its ad. “Close collaboration and partnership with the developer community allowed us to optimize these models for tasks that most count to their applications.”
In addition to performance gains, new models are delivered with a price reduction. GPT – 4.1 is 26% cheaper than GPT – 4O for median requests, and GPT – 4.1 Nano is the fastest and most affordable model of Openai to date, depending on the company.
The new models are now available for developers via the Openai API. With GPT – 4.1 offering better performance at a lower cost, Openai announced that it would stop supporting the GPT -4.5 overview on July 14.
Share this article
