OpenAI brings GPT-4o mini AI model targeting app developers: Check details

GPT-4o mini is now replacing the old GPT-3.5 model on OpenAI's chatbot - ChatGPT for Plus and Team subscribers, as well as free tier users. Enterprise customers to get access starting next week

OpenAI, on July 18, launched its new small GPT-4o mini AI model, branding it as their most cost-efficient offering. Targeting app developers with its new model, OpenAI in its press note stated that it expects that the GPT-4o mini will expand the range of applications built with AI, as the new model makes AI capabilities more affordable.

GPT-4o mini: Details

The Microsoft-backed AI startup company said that the GPT-4o mini surpasses GPT-3.5 Turbo model and other small models from other companies on benchmark scores across textual and multimodal reasoning. Additionally, it stated that the GPT-4o mini enables a variety of tasks with low latency such as passing a large volume of context to the model, interacting with customers with real-time text responses and more. 

OpenAI said that the GPT-4o mini has a context window of 128,000 tokens and supports up to 16,000 output tokens per request. The context window essentially determines the size of information an AI model can process. 

GPT-4o mini does not have access to real-time data, but the company said that it has knowledge up to October 2023. Additionally, the new model also supports the same range of languages as the GPT-4o model which is currently OpenAI’s most capable large language model (LLM).

GPT-4o mini: Availability

For consumers-

OpenAI has already started rolling-out GPT-4o mini in ChatGPT for Plus and Team subscribers as well as free tier users, replacing the GPT-3,5 model. Enterprise users of ChatGPT will get access to the GPT-4o mini starting next week.

For developers-

GPT-4o mini is also available as a text and vision model in the Assistants Application Programing Interface (API), Chat Completions API, and Batch API. Developers will have to pay 15 cents per 1 million input tokens and 60 cents per 1 million output tokens for using the new model.