Skip to content

最新のChatGPT-4o miniを新たに搭載!ファッション業界向け生成AIツール「Maison AI」が更に便利に。 ーGPT-3.5 Turboよりも高性能、画像解析搭載で生成速度とコスト効率に優れた小型モデルー

OpenFashion Co., Ltd. (Headquarters: Minato-ku, Tokyo, Representative: Toru Ueda, hereinafter referred to as OpenFashion) is pleased to announce that the latest ChatGPT-4o mini model has been newly installed in "Maison AI", a generative AI utilization support tool specialized for the fashion industry. This makes Maison AI a more effective and efficient AI generative tool.

OpenFashion Co., Ltd. is developing services and products that combine fashion with the latest technology, including AI. The company's "Maison AI" is a text and image generation AI tool specialized for the fashion industry, and is already being used by many people in a variety of situations, including improving business efficiency at companies, creating new businesses, and teaching at fashion schools.

◆Transition to ChatGPT-4o mini

With this update, Maison AI has transitioned its standard text chat model from GPT-3.5 turbo to ChatGPT-4o mini. GPT-4o mini outperforms previous models in accuracy and speed. In particular, the transition to ChatGPT-4o mini has enabled the upload of images as a new feature, and the number of characters that can be sent in a chat at one time has increased from 6,000 characters to 40,000 characters, about 6.6 times the previous number. This allows users to flexibly use AI to meet various needs, along with the already installed high-performance model GPT-4o.

◆ Performance comparison of GPT models

The graph below compares the performance of each AI model (GPT-3.5, GPT-4o mini, GPT-4o). As you can see from the graph, GPT-4o mini (orange bar) outperforms GPT-3.5 turbo (blue bar) in various areas.

Graph source: GPT-4o mini: advancing cost-efficient intelligence (https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/)

MMLU: Modeled language comprehension for a range of subjects including math, history and science.

GPQA: General Practical Question Answering Skills.

DROP: Ability to read whole paragraphs and engage in numerical and discrete reasoning.

MGSM: Ability to solve elementary school level mathematics problems.

MATH: Ability to solve problems across a range of mathematics including algebra, geometry, and calculus.

HumanEval: Ability to generate program code.

MMMU: Ability to understand tasks involving multiple data formats such as text, images, and audio.

MathVista: A facility that includes visual math problems and analysis.

With the introduction of GPT-4o mini, MaisonAI will no longer provide GPT-3.5, but you can continue to use GPT-4o mini without any restrictions. From now on, you can choose between the unlimited use of GPT-4o mini and the more accurate GPT-4o depending on your needs.

OpenFashion not only provides Maison AI to companies, but has also been supporting the promotion of its introduction so that the use of generative AI can spread among companies, especially in the fashion industry. With this update, we hope that more people will be able to experience generative AI, experience its potential firsthand, and enjoy it at the same time.

As a pioneer of technological innovation in the fashion industry, we will continue to contribute to the sustainable development of the industry by utilizing generative AI.

For more information about this and past updates, please see below.

https://maisonai.io/blogs/update/ver1-19

About the generative AI utilization support tool "MaisonAI"

■MaisonAI service site (new registrations receive a 30-day free trial)

https://maisonai.io/


■MaisonAI Guide Page (notion)

You can find out more about how to use MaisonAI here.

https://omnis.notion.site/5869e946b86e4c9191ce729bc7751db2


■ MaisonAI update information

You can see the latest and past updates here

https://maisonai.io/blogs/update

◆ "Maison AI" Service Overview

"Maison AI" provided by OpenFashion Co., Ltd. is a generative AI support tool specialized for the fashion industry. Maison AI supports the latest ChatGPT model GPT-4o and demonstrates high performance in both language understanding and image analysis. It provides a multi-functional toolset including AI Agent Pro, and is used in a variety of situations, including business efficiency, new business creation, and educational purposes.

Main Features

1. Supports GPT-4o : The accuracy of language understanding and response speed have been improved, enabling faster and more accurate information to be provided.

2) Image analysis : Image content description, character recognition, and tagging are performed automatically, dramatically improving the efficiency of marketing and product development.

3) AI Agent Pro : AI agents specialized in each profession provide powerful support to users' work.

④AI Design : You can generate various design images by inputting text.