Photo credit: www.theverge.com

OpenAI Unveils GPT-4.1, Enhancing AI Capabilities

OpenAI has officially launched GPT-4.1, the successor to the previously released multimodal AI model GPT-4o. In a livestream event on Monday, the company highlighted that GPT-4.1 boasts a significantly expanded context window and outperforms GPT-40 across nearly all aspects, particularly in coding tasks and following instructions.

For developers, GPT-4.1 is readily available, accompanied by two additional, smaller model variants. The GPT-4.1 Mini model is designed to be more accessible and cost-effective for experimentation, while the GPT-4.1 Nano is touted as the smallest, fastest, and most economical model released to date. Each of these models is capable of handling up to one million tokens in context, which includes the data types such as text, images, or videos provided in a single prompt.

This launch comes as OpenAI prepares to phase out its older GPT-4 model from ChatGPT, effective April 30th. In a changelog announcement, the company described recent enhancements to GPT-4o as making it a “natural successor.” Additionally, OpenAI plans to retire the GPT-4.5 preview from its API on July 14th, citing that “GPT-4.1 delivers improved or comparable performance on numerous critical functionalities at a reduced cost and latency.”

Moreover, OpenAI is on the brink of releasing the full version of its o3 reasoning model, alongside an o4 mini reasoning model. Early references to these models have already been discovered in the latest ChatGPT web update, as noted by AI engineer Tibor Blaho on social media.

Source
www.theverge.com