OpenAI introduced GPT-4.1 and GPT-4.1 mini to enhance coding assistance in ChatGPT for software engineers, with subscriptions for Plus, Pro, and Team users now available. This update replaces the previous GPT-4.0 mini model. Notably, while GPT-4.1 shows improved performance, it does not meet the criteria for a frontier model, exempting it from stringent safety report requirements. OpenAI aims to enhance transparency by committing to publish internal safety evaluations regularly through a new hub dedicated to safety information.
OpenAI is now rolling out GPT-4.1 to subscribers of ChatGPT Plus, Pro, and Team, while free and paying users will have access to GPT-4.1 mini.
Despite improved performance, OpenAI clarified GPT-4.1 was not classified as a frontier model, thus exemptions from full safety reporting apply.
OpenAI's Head of Safety Systems stated that although GPT-4.1's safety considerations are substantial, they differ from those of frontier models.
OpenAI announced plans to increase transparency by publishing internal safety evaluations more frequently and launched a new Safety Evaluations Hub.
Collection
[
|
...
]