GPT-4.1 Mini Replaces GPT-4o Mini as Default Model for All ChatGPT Users

On a post on X on 14 May, OpenAI declared that its GPT-4.1 and GPT-4.1 mini AI models would be available on ChatGPT.

Software engineers utilising ChatGPT should find the GPT-4.1 models useful for writing or debugging code, according to OpenAI spokesman Shaokyi Amdo, who spoke to a media house.

OpenAI claims that GPT-4.1 is faster than its o-series of reasoning models yet performs better at coding and command following than GPT-4o. According to the firm, ChatGPT Plus, Pro, and Team subscribers are currently receiving GPT-4.1.

 In the meantime, ChatGPT customers that pay can download GPT-4.1 mini for free from OpenAI. According to the company's release notes for GPT-4.1, OpenAI is eliminating GPT-4.0 small from ChatGPT for all users as a result of this update.

Open AI Facing Criticism

GPT-4.1 and GPT-4.1 mini were introduced by OpenAI in April, although the models were only made available via its developer-facing API. The AI research community at the time criticised the corporation for shipping GPT-4.1 without a safety report.

According to these academics, OpenAI was lowering the bar for transparency in their AI models. At the time, OpenAI contended that even while GPT-4.1 outperformed GPT-4o in terms of speed and performance, it was not a frontier model and, as a result, did not need the same safety reporting as more advanced models.

In a post on X, Johannes Heidecke, Head of Safety Systems at OpenAI, claimed that GPT-4.1 does not outsmart O3 and does not add any new modalities or ways to engage with the model. This implies that, although significant, the safety issues in this context differ from those in frontier models, he continued.

In contrast to GPT-4o's 128,000-token restriction, both of the new models offer a one million context token window, which is the maximum amount of text, images, or videos in a prompt that an AI model can analyse. According to OpenAI, speed enhancements also make GPT-4.1 more desirable for routine coding activities.

Open AI Clearing the Air With Live Demonstration

More details on GPT-4.1 and all of its AI models are now being made public by OpenAI. In an attempt to be more transparent, OpenAI earlier on 14 May pledged to release the findings of its internal AI model safety assessments more regularly.

OpenAI inaugurated its new Safety Evaluations Hub on Wednesday, and those findings will be available there. The launch of ChatGPT's GPT-4.1 coincides with a surge in interest in AI coding tools.

One of the most well-known AI coding tools available, Windsurf, is apparently about to be acquired by OpenAI for $3 billion. Google modified its Gemini chatbot earlier to make it easier to connect to GitHub projects.