On July 18, US-based artificial intelligence firm OpenAI announced the release of a new generative AI model called “GPT-4o Mini.” This new model is touted to offer performance comparable to its predecessors at a significantly reduced cost.

According to a blog post from OpenAI, GPT-4o Mini is “an order of magnitude more affordable than previous frontier models,” and “more than 60% cheaper than GPT-3.5 Turbo.”

Balancing Performance and Power

GPT-4o Mini serves as a cost-effective version of OpenAI’s top consumer model for ChatGPT. Despite its reduced energy consumption, the model reportedly maintains high performance levels. OpenAI stated that GPT-4o Mini surpasses GPT-3.5 Turbo and other smaller models in academic benchmarks across both textual intelligence and multimodal reasoning. It also supports the same range of languages as GPT-4o.

Features and Functionality

The new model retains much of the functionality of its predecessor, currently supporting text and vision. OpenAI plans to add support for audio and video soon. However, it remains unclear if GPT-4o Mini offers any environmental benefits. OpenAI has not disclosed the specific methods used to reduce running costs, which might suggest that the savings are primarily financial rather than energy-related.

Ongoing Developments at OpenAI

The launch of GPT-4o Mini is part of a series of recent activities by OpenAI. As reported by Cointelegraph, OpenAI is allegedly working on another AI model, codenamed “Strawberry,” which is expected to offer advanced reasoning capabilities beyond GPT-4o, delivering more humanlike responses. Strawberry is believed to be an extension of OpenAI’s enigmatic Q* project.

Simultaneously, OpenAI could face scrutiny from the Securities and Exchange Commission following whistleblower allegations regarding the company’s use of non-disclosure agreements.

Source: Cointelegraph