OpenAI Launches GPT-6.1 Terra, AI Can Control Your Computer and Automate Tasks

New Delhi: The world's leading artificial intelligence company OpenAI has launched GPT-6.1 Terra. This new model of the company is not limited to just answering questions or writing text but also has the ability to work in various apps by understanding the computer screen. It has been specially designed for tasks that require repetitive information entry, filling out forms or coordination between different software. According to the company, this model can help in professional work with better visual reasoning, faster functioning and lower cost.

How will GPT-6.1 Terra work on the computer itself?

The speciality of GPT-6.1 Terra is its ability to use computers. It can understand the information displayed on the screen and perform multi-step tasks in apps and browsers using available tools. It can be used for tasks like filling online forms, extracting information from documents, preparing proposals and entering data into different software. However, for this, setup of related tools and facilities to use computer is necessary. That is, it is not a system that runs the computer automatically in every situation without any permission or arrangement.

How will businesses and companies benefit?

This model can be useful for companies in those tasks in which large amounts of information have to be transported from one place to another. For example, extracting information from bills received from suppliers and entering them into accounting software, understanding technical drawings or filling out long digital forms. OpenAI also includes complex tasks like creating 3D models from technical drawings and reviewing videos as examples of its computer-use tests. This can save employees time on repetitive tasks.

How much faster and cheaper is the new version than the old model?

According to OpenAI, GPT-6.1 Terra outperformed GPT-5.6 Terra in the company's computer-usage test Agents' Last Exam. Its score on the intermediate reasoning level was 15 percentage points higher than the old model. The company claims that it can complete certain tasks on average more than three times faster than GPT-6 Astra and GPT-6.1 Sol. These results are based on company tests and actual performance may vary depending on the nature of the work and the tools used.

What is special about mistakes and coding?

In OpenAI's internal testing, GPT-5.6 Terra's error rate on difficult questions at the low reasoning level was 53.6 percent, compared to 16.7 percent for GPT-6.1 Terra. The new model's error rate at the maximum reasoning level was 4.3 percent. It also solved 50.3 percent of tasks at the maximum reasoning level in the Terminal-Bench 4.0 coding test, compared to the older GPT-5.6 Terra's published score of 21.5 percent. These figures are from the company's tests and should not be considered a guarantee of error-free performance in every actual operation.

What is the price and availability of GPT-6.1 Terra?

According to OpenAI's official announcement, GPT-6.1 is available for Plus, Pro, Enterprise, and Edu users in Terra ChatGPT Work and Codex. Developers can also use it through OpenAI API. Its standard API price is one dollar per million input tokens, $0.05 per cached input token, and five dollars per output token. According to the company, its standard API prices are 50 percent lower than GPT-5.6 Terra. This offers the potential to reduce costs for developers and businesses looking to automate large numbers of digital tasks.