Financial AI "arms race" escalates! OpenAI launches a full suite of service tools to compete with Anthropic for Wall Street clients.

OpenAI has released a new flagship artificial intelligence model and a set of financial service tools aimed at better handling office work, intensifying competition with Anthropic, which faces new risks from its confrontation with the Pentagon. OpenAI announced on Thursday that the GPT-5.4 model can better handle tasks such as generating spreadsheets, documents, and presentations, significantly reducing the need for users to communicate repeatedly with it. The model is designed to excel in finding answers to challenging problems by accessing information from multiple sources, analyzing it, and summarizing responses. The company also announced the launch of a new set of tools designed to help professionals simplify financial analysis, investment memos, and other work. This product can connect to financial data and research companies such as FactSet Research Systems Inc. and Third Bridge using the ChatGPT application. OpenAI stated that users can now directly use ChatGPT in Excel and Google Sheets to create and review financial models. Business competition and pressure from high valuations OpenAI and Anthropic have been trying to convince more business professionals to pay for their services to offset the high costs of developing artificial intelligence systems and support their high valuations. Anthropic has particularly emphasized support for financial services professionals in recent product releases; the company also launched "Claude Financial Service Edition" last year. The increasingly fierce competition between these two companies comes at a time when Anthropic's business is in an uncertain period. Previously, after the Pentagon announced "supply chain risks" for Anthropic following a request for protective measures to ensure its technology was not used for mass surveillance of Americans or deploying fully autonomous weapons, OpenAI reached an agreement allowing the Pentagon to deploy the company's AI model in its classified network within hours. OpenAI CEO Sam Altman later stated that OpenAI's eagerness to strike a deal with the Department of Defense appeared "opportunistic and hasty." He also mentioned that the company was working with the department to "add some content to our agreement to make our principles very clear." Meanwhile, Anthropic has resumed negotiations with the Pentagon. OpenAI, Anthropic, and other AI developers have released multiple models in recent months, claiming that they can represent individuals in completing a wider range of tasks and further expand on the progress made by tools that previously simplified the writing and debugging of code. This momentum is causing concern among investors in traditional software companies, who fear that these companies may be displaced by AI enterprises. OpenAI is introducing its new model called "GPT-5.4 Thinking" to paid subscription users, indicating that it will spend more time deeply analyzing responses to prompts. The model will also be available through the company's AI code assistance tool Codex. GPT-5.4 Thinking will demonstrate its thinking process in handling complex tasks, allowing users to express opinions or adjust direction as needed during the response generation process.