OpenAI has launched a new series of models known as GPT-4.1, which includes the GPT-4.1 mini and nano, claimed to outperform the earlier GPT-4o models at a lower cost. With features like larger context windows of one million tokens and increased output limits, these models are designed to handle complex tasks and improve long-context comprehension. The GPT-4.1 models will only be accessible via the API, as OpenAI phases out the GPT-4.5 Preview by July 2025. Notably, the mini variant shows a considerable leap in performance, especially for coding tasks, which has garnered positive feedback from the developer community.
OpenAI says that its new models, GPT-4.1, and its variants mini and nano, outperform the older GPT-4o series while being more affordable.
The introduction of the GPT-4.1 series brings significant improvements in context window sizes and output token limits, greatly enhancing long-context comprehension.
With a 21.4% improvement in coding performance over GPT-4o, particularly noticeable in the GPT-4.1 mini model, OpenAI emphasizes advancements for developers.
GPT-4.1 nano is highlighted as exceptionally fast and cost-effective, providing remarkable performance with a 1 million token context window.
Collection
[
|
...
]