OpenAI Launches New Coding-Focused GPT-4.1 Models

OpenAI contiguous announced the motorboat of three caller GPT models that are disposable done the OpenAI API. Called GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano, the models are not disposable for ChatGPT astatine the existent time.


GPT-4.1 includes large improvements successful coding, acquisition following, and agelong context, according to OpenAI. The models outperform GPT-4o and GPT-4o mini successful each tasks, particularly coding tasks. The caller models enactment up to 1 cardinal tokens of context, and tin usage that for amended long-context comprehension.

On SWE-bench Verified, GPT-4.1 showed a 21.4 percent betterment implicit GPT-4o for coding, and a 26.6 percent betterment implicit GPT-4.5. The slimmed down GPT-4.1 mini matches oregon exceeds GPT-4o with reduced latency and an 83 percent little cost. GPT-4.1 nano is OpenAI's fastest and cheapest model, perfect for classification, autocompletion, and akin tasks.

OpenAI says that galore of the improvements successful GPT-4.1 person already been incorporated into the ChatGPT mentation of GPT-4o, with much functionality to beryllium added successful the future. All GPT-4.1 models person a cutoff day of June 2024, which means they are up to day connected events that took spot earlier then.

With the instauration of GPT-4.1, OpenAI is deprecating GPT-4.5 successful the API, arsenic GPT-4.1 offers overmuch of the aforesaid functionality astatine a little cost.

GPT-4.1 costs $2 per cardinal input tokens and $8 per cardinal output tokens, portion GPT-4.1 mini costs $0.40 per cardinal input tokens and $1.60 per cardinal output tokens. GPT-4.1 nano is conscionable $0.10 per cardinal input tokens and $0.40 per cardinal output tokens. Customized models are disposable astatine a higher price.

This article, "OpenAI Launches New Coding-Focused GPT-4.1 Models" archetypal appeared connected MacRumors.com

Discuss this article successful our forums