OpenAI unveils ChatGPT Work, its new agentic 'super app', and intensifies the rivalry with Anthropic

OpenAI on Thursday, July 9, 2026, unveiled a new AI agent called ChatGPT Work, designed to let office workers without programming skills tap into the power of coding tools without taking on their usual cost. According to the company, the agent combines its popular chatbot ChatGPT with its coding tool Codex to create documents, presentations and websites. The launch comes alongside GPT-5.6, described as OpenAI's most advanced model to date, which debuted the same day after its release was delayed the previous month at the request of the U.S. government on national security grounds.

The move is a direct response to Claude Cowork, the agent Anthropic launched in January that is capable of planning and executing multi-step tasks autonomously. Reuters frames this launch within an increasingly intense competition among the big tech companies to build and sell AI tools for professional use, at a time when demand for autonomous agents capable of completing complex tasks with minimal human intervention is on the rise. Both Anthropic and OpenAI, according to the article, are preparing for possible stock market debuts soon and are competing hard for enterprise business, which is more lucrative than selling products to consumers.

A key point of the announcement is that both ChatGPT Work and Claude Cowork are designed for the same audience: users who do not program but want access to the capabilities of coding tools, which are usually more powerful than standard chatbots but require specialized knowledge. OpenAI executives insisted that their new offerings will be cheaper and more widely available than those of their rivals. The company, in fact, launched three different sizes of the 5.6 model. Ty Geri, product manager for ChatGPT Work, said in an interview that "you can apply the model's coding ability to solve problems in any industry", and described the 5.6 model as "competitive with models that are far more expensive, at twice the speed and much, much cheaper".

The article also captures a growing concern among companies about the high cost of using this type of AI tool. Max Weinbach, an analyst at the consulting firm Creative Strategies, noted that the smallest version of OpenAI's new model completes a task nearly as well as the largest version, but at one-fifth of the cost. "This is the first time I've seen small models complete this kind of task", he said.

As for the product context, Reuters recalls that before this launch OpenAI's agentic offering included Operator and deep research, later consolidated into ChatGPT Agent for individual users, in addition to Workspace Agents for automating enterprise workflows. ChatGPT Work rolls out from Thursday on web and mobile, starting with users on the Pro, Enterprise and Edu plans, and will extend to Plus and Business users in the coming days. Alongside this agent, OpenAI also announced a new ChatGPT desktop app and a hosted websites feature that allows web pages to be created and shared directly from ChatGPT Work.

Taken together, the news confirms that the dispute between OpenAI and Anthropic has shifted from the terrain of foundation models to that of agents geared toward office productivity, with Microsoft Copilot Cowork also mentioned as a direct competitor in that same space. OpenAI's emphasis on the price and efficiency of smaller models suggests that the next phase of this rivalry will be fought over both capabilities and access costs for companies, a factor that, according to the article itself, has become a central concern for corporate customers.