KMS Updates Business Efficiency Service Utilizing AI Agents, Newly Implementing 5 Latest AI Models Including "GPT-5.4" from Azure AI Foundry

KMS Corporation (President and Representative Director: Kentaro Kajiwara), which supports business efficiency utilizing AI Agents, has announced the implementation of new AI models "GPT-5.3-Chat", "GPT-5.4", "GPT-5.4-Pro", "GPT-5.4-mini", and "GPT-5.4-nano" into "KMS's AI Solution" leveraging Azure AI Foundry.

With this update, the accuracy of understanding complex instructions and context has been further improved, making it possible to support tasks such as text creation, data analysis, and proposal writing at a more advanced and efficient level. By equipping the service with advanced conversational understanding and multifaceted analytical capabilities, it realizes more natural, consistent, and highly accurate output compared to conventional models.

Additionally, because users can select the optimal model according to their applications and business requirements, it can flexibly accommodate a wide range of use cases, from advanced operations emphasizing accuracy to high-volume processing prioritizing cost and speed.

"KMS's AI Solution" will further expand the utilization of AI in various business areas such as text creation, image generation, and analysis support, and will strongly support corporate business efficiency and DX (Digital Transformation) promotion.

* You can download use case materials here

■ What is KMS's AI Solution?

"KMS's AI Solution" allows the utilization of RAG (Retrieval-Augmented Generation) under a secure environment. It enables maximum utilization of the data held by customers, further boosting business efficiency. Furthermore, by making three elements interact—User Interface (creative power), Development (development and operational power), and User Experience (experience)—we are challenging new technologies. In addition, by enabling use on Microsoft Teams (hereinafter "Teams") and connecting with external services like Box and SharePoint, we provide an unprecedented experience.

* RAG is an AI technology that combines information retrieval and generation. In this method, the AI Agent quickly searches for relevant information from massive databases and documents, and generates new content and answers based on the obtained information.

---

■ Additional Models: GPT-5.3-Chat, GPT-5.4, GPT-5.4-Pro, GPT-5.4-mini, GPT-5.4-nano

▼ GPT-5.3-Chat (A versatile conversational model at the core of business efficiency)

A highly versatile model that combines advanced natural language understanding with stable conversational performance. It has strengths in summarizing and structuring long texts and creating business documents, with the major advantage of easily providing output of a consistent quality. It is particularly suitable for searching and summarizing internal knowledge and automating inquiry responses, achieving both accuracy and reproducibility when combined with a RAG configuration. It is a model that realizes wide-ranging business efficiency in customer support, internal FAQs, minute generation, and more.

---

▼ GPT-5.4 (An advanced reasoning model that accelerates decision-making)

An upper-tier model with enhanced reasoning capabilities and context understanding, capable of handling complex business logic and advanced analytical tasks. Its advantage lies in its ability to generate consistent output even for highly abstract instructions, making it possible to improve the accuracy of organizing thoughts and supporting judgments. It is effective for assisting consulting operations, creating proposal outlines, and interpreting data analysis, improving the speed and quality of decision-making simultaneously. It is a model that directly leads to productivity improvement in the planning and strategy domains.

---

▼ GPT-5.4-Pro (The top-tier model that responds to fields requiring high accuracy)

The highest-level model equipped with the highest standard of accuracy and reasoning power. Its greatest advantage is the ability to provide stable, high-quality output in highly specialized fields and operations where mistakes are unforgivable. It is ideal for uses demanding high precision, such as legal checks, contract reviews, and complex system design support. It is also strong in problem-solving that requires multi-stage thinking, demonstrating its value in projects and enterprise applications where quality is the top priority.

---

▼ GPT-5.4-mini (A practical balance model that achieves both cost and speed)

A balanced model characterized by its lightweight and high-speed response. Its advantage is keeping costs low while maintaining a certain level of accuracy, making it ideal for processing high-volume requests. It delivers high performance in primary chatbot responses, simple text generation, and classification/tagging. It is suitable for situations where speed and cost efficiency are important, and is effective for services with many user touchpoints and systems that require real-time processing.

---

▼ GPT-5.4-nano (An ultra-lightweight model that realizes overwhelming scalability)

A model realizing ultra-lightweight and ultra-high speed, capable of processing simple tasks at a low cost. Its advantage is enabling large-scale operation while minimizing processing load, excelling in scalability. It is strong in short sentence responses, standard text generation, and log classification, making it suitable for IoT, embedded environments, and services with many simultaneous requests. It demonstrates its effectiveness in use cases where speed and cost are prioritized, such as simple chat responses and input assistance.

---

■ [Azure AI Foundry]

By leveraging cutting-edge AI models such as "GPT-5.3-codex", "GPT-5.3-chat", "GPT-5.4", "GPT-image-1.5", "Deep Research", "Grok 4", and "Anthropic Claude Sonnet/Opus", it provides a custom experience that flexibly handles diverse tasks like natural conversation, text generation, code creation, and image generation.

■ Company Profile

Company Name: KMS Corporation (Kaleidoscope Media Service)
Representative: Kentaro Kajiwara, President and Representative Director
Location: 15F Tennoz First Tower, 2-2-4 Higashi-Shinagawa, Shinagawa-ku, Tokyo
Established: July 2015
Capital: 29,980,000 yen
Business Description: AI business, game business, cloud solution business, digital comic business
Corporate Website: https://kms3.com/

KMS's AI Solution: https://daiverse.jp/

* Microsoft Azure is a registered trademark or trademark of Microsoft Corporation in the United States and other countries.
* Other company names and product names mentioned are trademarks or registered trademarks of their respective companies.