GuShiio.com Article source: Interface News
Image source: Generated by AI
On March 6, The Information reported that OpenAI plans to charge AI Agents that reach doctoral level US$20,000 (approximately RMB 145,000) per month, mainly for the high-end needs of corporate users, especially in finance, medical care, manufacturing, etc. Data-intensive industries.
It is reported that these “doctoral level” AI Agents can handle complex tasks in the fields of academic research and software development. OpenAI plans to launch three types of AI Agents, with monthly charges likely ranging from $2000 to $20,000.
At the end of last year, OpenAI spoilt relevant information in the live broadcast. In addition to the regular version upgrade of GPT-4o, the mysterious “Agents” folder was seen as implying that “AI Agents” were coming soon.
The CFO of OpenAI also revealed in an interview that the company is planning to launch an AI Agent that can perform complex tasks. This AI Agent will have a doctoral level and be able to help users in various matters, thereby creating economic value for the company. Therefore, its pricing model will no longer be simply based on the number of tokens or the number of heads used, but will shift to pricing based on the value the product can bring. As soon as the news came out, it quickly attracted widespread attention and heated discussion.
Faced with high subscription fees, some people joked that their salary was not as high as ChatGPT, and some company bosses even said that they would rather hire a real doctor at this price.
However, OpenAI clearly targets large enterprises and industry giants rather than individual users. For corporate customers, OpenAI also plans to charge a fixed fee based on employee seats and can charge customers based on the value they get from the product, which will help offset OpenAI’s huge cost of developing AI systems.
No matter how companies now define “agent”, the core is to complete the given work with as little human interaction as possible.
Compared with chatbots that can only ask and answer questions,”AI agents” are software used to handle specific tasks, usually requiring multiple steps and calling different tools. For example, the task of booking tickets requires AI to search online to check ticket website information, select the most suitable flight and call payment software to purchase tickets.
Almost all technology companies on the AI track are eyeing “AI agents” because they are products that are easier to sell at a price. Google released its first AI proxy tool, Project Mariner, at the end of last year, which can be used to find flights, hotels and other tasks.
Because AI can work without interruption, if the work output is also satisfactory, technology companies will be more able to persuade more companies to use AI “employees.”
However, Dario Amodai, co-founder of OpenAI’s competitor Anthropic, also put forward a reflective view. He believes that the widespread use of AI agents may bring higher rewards to employees with skills that AI cannot replicate.
