OpenAI Launches GPT-6.1 Terra, Delivering Cheaper AI Automation and Improved Performance

OpenAI has officially launched GPT-6.1 Terra, a new artificial intelligence model designed to make computer-based automation faster, more efficient and less expensive. Announced on October 26, 2026, the model represents an upgrade over GPT-5.6 Terra, with improvements in visual reasoning, computer interaction and professional writing. Unlike AI systems primarily designed to answer questions or generate text, GPT-6.1 Terra focuses on completing practical tasks across digital applications. These include drafting business proposals, interpreting technical drawings, filling out online forms, processing invoices and editing videos. The release reflects OpenAI’s broader strategy of developing AI agents that can do more than provide instructions by actively carrying out tasks on a computer. The company says the model offers a better balance between speed, capability and operating costs, potentially making advanced automation accessible to more businesses and developers.

One of the model’s most important capabilities is its ability to interpret what appears on a computer screen and take actions across different applications. This allows it to follow multistep instructions, analyse visual information and perform work that would traditionally require a person to navigate software manually. For example, an AI agent could read information from incoming invoices and enter the relevant details into accounting software, helping businesses reduce repetitive administrative work. OpenAI also reports improvements in visual reasoning and desktop navigation compared with its previous Terra model. On the company’s Agents’ Last Exam benchmark, GPT-6.1 Terra scored 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort while costing approximately one-third as much for the evaluated tasks. These results are based on OpenAI’s reported testing, and independent evaluations will help establish how consistently the model performs in real-world environments.

Accuracy is another major area of improvement. OpenAI says GPT-6.1 Terra produced fewer factual errors than its predecessor when tested against prompts designed to expose mistakes. At low reasoning effort, the reported error rate declined from 53.6% with GPT-5.6 Terra to 16.7% with the new model. At maximum reasoning effort, the company reported an error rate of 4.3%. The model also demonstrated stronger coding performance, solving 50.3% of tasks on the Terminal-Bench 4.0 benchmark at maximum reasoning effort, compared with 21.5% for GPT-5.6 Terra. These improvements could make the model more useful for software developers, technical teams and businesses seeking to automate routine digital workflows. However, lower error rates on specific benchmarks do not mean the model is error-free, and organisations should continue checking important outputs before using them for financial, legal or operational decisions.

Cost is central to GPT-6.1 Terra’s appeal. OpenAI has set its standard API pricing at $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens. The company says these rates are 50% lower than those of GPT-5.6 Terra and GPT-6.1 Sol at standard API prices. The model is rolling out to Plus, Pro, Enterprise and Edu subscribers through ChatGPT Work and Codex, while developers can access it through the API using the model identifier gpt-6.1-terra. Its lower operating costs could benefit companies processing large volumes of documents, building customer-support systems or developing AI-powered business applications. Nevertheless, the total cost of an automated workflow will also depend on how much reasoning, computer interaction and repeated processing each task requires. Businesses should test the model against their specific workloads before deciding whether it offers better value than alternative systems.

Ultimately, the launch of GPT-6.1 Terra highlights the growing importance of AI agents in the technology industry. Competition is increasingly focused not only on which model can answer the most difficult questions, but also on which system can complete useful work accurately, quickly and affordably. If these capabilities continue improving, AI agents could transform administrative operations, software development, research assistance and other computer-based professions. At the same time, companies must address privacy, security, reliability and human oversight when allowing AI systems to interact with important applications. OpenAI’s latest release represents another step towards practical AI automation, although its long-term impact will depend on performance outside controlled tests and the ability of businesses to integrate it safely. As developers begin experimenting with GPT-6.1 Terra, the coming months should reveal how effectively its lower prices and improved computer-use capabilities translate into measurable productivity gains.