Post Thumbnail

OpenAI introduced GPT-4.1 — a new leader in programming and task following

OpenAI has released a new version of its flagship model called GPT-4.1. It is positioned as a specialized model focused on performing programming tasks and precise instruction following. GPT-4.1 is already available in ChatGPT. If you don’t see it – log out and log back into your account. In addition to the main version, OpenAI also introduced GPT-4.1 mini, which replaces the previous GPT-4 o mini.

The key advantage of the new model lies in its high performance when working with program code. In the industry benchmark SWE-bench Verified, the model demonstrates a result of 54.6%. This surpasses the performance of GPT-4 o by 21.4% and GPT-4.5 by 26.6% points. Making it the leading model for coding tasks among available artificial intelligence systems.

In instruction following tests conducted using Scale’s MultiChallenge methodology, the new GPT-4.1 reached 38.3%. Which is 10.5% better than the results of GPT-4 o. The model also sets a new record in understanding long multimodal context according to the Video-MME benchmark. Scoring 72.0% in the category of long videos without subtitles. Which is 6.7 percentage points better than the previous version.

Special attention should be paid to working with large volumes of data. GPT-4.1 supports context up to 1 million tokens through the application programming interface. However, for ChatGPT users, the limitations remain the same. 32 thousand tokens for Plus subscription and 128 thousand for Pro subscription.

OpenAI developers note that when creating the model, they focused not only on synthetic benchmarks but also on practical usefulness for real-world tasks.

Autor: AIvengo
For 5 years I have been working with machine learning and artificial intelligence. And this field never ceases to amaze, inspire and interest me.

Latest News

The first LAARMA system protects animals on Australian roads

In Australia, animal-vehicle collisions are a serious problem for this continent's ecosystem. Now scientists have found a technological solution. The world's first roadside LAARMA system based on artificial intelligence that protects wild animals from dangerous encounters with traffic.

Nvidia introduced Cosmos model family for robotics

Nvidia company introduced the Cosmos family of AI models. Which can fundamentally change the approach to creating robots and physical AI agents.

ChatGPT calls users "star seeds" from planet Lyra

It turns out ChatGPT can draw users into the world of scientifically unfounded and mystical theories.

AI music triggers stronger emotions than human music

Have you ever wondered why one melody gives you goosebumps while another leaves you indifferent? Scientists discovered something interesting. Music created by artificial intelligence triggers more intense emotional reactions in people than compositions written by humans.

GPT-5 was hacked in 24 hours

2 independent research companies NeuralTrust and SPLX discovered critical vulnerabilities in the security system of the new model just 24 hours after GPT-5's release. For comparison, Grok-4 was hacked in 2 days, making the GPT-5 case even more alarming.