Adapted by:
Javier Márquez
WriterI've been in media for over a decade, but I've been marveling at the possibilities that technology brings us much longer. I believe we live in a world where the digital revolution is changing everything and that Xataka is the best place to write about it.
Alba Mora
WriterAn established tech journalist, I entered the world of consumer tech by chance in 2018. In my writing and translating career, I've also covered a diverse range of topics, including entertainment, travel, science, and the economy.
Until recently, GPT-4.1 was exclusively available to developers, accessible only through the OpenAI API. This meant that regular ChatGPT users couldn’t use it. However, this is about to change.
GPT-4.1 is coming to ChatGPT. OpenAI has begun rolling out the GPT-4.1 model within its chatbot service for all paying subscribers, including ChatGPT Plus, Pro, and Team. For Enterprise and Edu accounts, the company will provide access in the coming weeks.
What about free users? Currently, free ChatGPT users don’t have access to the full version. Instead, OpenAI is introducing GPT-4.1 mini, a lighter variant with fewer capabilities. This version improves upon GPT-4o mini and is expected to be enough for many everyday tasks.
New model list. With this change, OpenAI has removed GPT-4o mini from the ChatGPT model selector and replaced it with GPT-4.1 mini. This new model serves as the starting point for most users, including those on the free plan.
Improved performance, larger context, and lower costs (at least for the API). One notable advancement in GPT-4.1 is its ability to process up to one million tokens. This is eight times the length of the full source code for React. Currently, this capability is only available through the API. It remains unclear whether it’ll be included in the ChatGPT version.
A more capable model. Benchmarks published by OpenAI highlight significant improvements across the board. In instruction-following tasks, GPT-4.1 achieved a score of 38.3% on the Scale’s MultiChallenge benchmark. This represents a 10.5-point increase compared to GPT -4.
When it comes to understanding long context, GPT-4.1 scores 72% in the “no subtitles” category of the Video-MME benchmark. It surpasses its predecessor by 6.7 points.
In coding, a focus area for the company, the model achieved a score of 54.6% on the SWE-bench Verified benchmark. It clearly outperforms both GPT-4 and GPT-4.5.
A genuine alternative to o3? GPT-4.1 isn’t specifically designed as a reasoning model. However, OpenAI claims that it can perform as well as or even better than o3 in development tasks. The startup’s goal is to combine the versatility of standard models with the precision of models tailored for reasoning.
Images | Solen Feyissa | Levart_Photographer
View 0 comments