Investing

OpenAI Unveils New GPT-4.1 Models, Phases Out Disappointing GPT-4.5 Preview

Harrison Brown 2 weeks ago 2 min read

OpenAI has introduced a new lineup of AI models aimed primarily at developers, while simultaneously announcing the discontinuation of a previous model that had failed to meet expectations.

During a livestream on Monday, the U.S.-based company revealed the new GPT-4.1 series, which includes the variants GPT-4.1 mini and GPT-4.1 nano. These models promise significantly improved capabilities, particularly in understanding and processing longer texts, as well as supporting developers with more effective coding assistance.

According to OpenAI, the updated models exhibit notable enhancements in generating computer programs, following user instructions with greater precision, and handling large amounts of information. Alongside the launch, OpenAI also announced it will be shutting down the API for its GPT-4.5 Preview model — a version that had drawn criticism from many industry experts for its underwhelming performance.

One of the most notable upgrades in GPT-4.1 is its expanded context window. With support for up to one million tokens — roughly 750,000 words — the new models can manage extensive and complex inputs more efficiently than before.

The models are also multimodal, meaning they can process both text and images. OpenAI stated that the GPT-4.1 models are trained with data up to June 2024, providing developers with more up-to-date knowledge.

In addition to performance improvements, OpenAI emphasized cost and speed advantages. GPT-4.1 is reportedly 26% cheaper than GPT-4.0, which remains the default model for most ChatGPT users. This reduction in cost is expected to make the models more accessible for a wider range of applications.

For developers seeking even faster and more efficient solutions, the mini and nano versions are optimized accordingly. However, OpenAI acknowledged that these versions may involve slight trade-offs in accuracy. Among the two, GPT-4.1 nano stands out as the fastest and most cost-effective model the company has ever released.

With these updates, OpenAI continues to position itself at the forefront of the generative AI space, aiming to provide flexible tools that meet the growing demands of developers, while also phasing out models that don’t meet evolving performance standards.