OpenAI Launches GPT-6.1 Terra With Advanced Computer Use Capabilities

Ender Öztürk

October 26, 2026

OpenAI officially announced the release of its latest artificial intelligence model, GPT-6.1 Terra, on October 26, 2026, marking a significant advancement in computer automation and visual reasoning. Designed to bridge the gap between human instruction and digital execution, the new model can interpret screen content, interact with various software applications, and perform complex workflows autonomously. GPT-6.1 Terra is now available to ChatGPT Plus, Team, Enterprise, and Edu subscribers, as well as developers through the OpenAI API. This launch aims to provide a more efficient and affordable solution for businesses seeking to automate routine computer-based tasks during the workday.

Computer Automation Becomes More Affordable for Developers

One of the most striking features of GPT-6.1 Terra is its aggressive pricing strategy, which makes high-volume automation more accessible than ever. OpenAI has set the standard API prices for this model at approximately 50 percent lower than both GPT-5.6 Terra and the more robust GPT-6.1 Sol. Specifically, the model costs 1 dollar per million input tokens, 0.05 dollars per million cached input tokens, and 5 dollars per million output tokens. This reduction is intended to encourage developers to build agents that can handle repetitive, high-volume tasks such as reading thousands of invoices and entering data into accounting software without human intervention.

Beyond just being cheaper, the model is designed to be significantly faster. OpenAI reports that GPT-6.1 Terra can complete tasks up to three times faster than GPT-6 Astra or GPT-6.1 Sol. When performing tasks at the highest reasoning level, Terra also demonstrates a remarkable cost advantage over its competitors. For instance, the company claims that the model provides nearly nine times lower costs per task compared to Anthropic’s Claude Opus 5.5. This combination of speed and affordability positions Terra as a primary choice for enterprise-level automation where efficiency is the most critical factor.

Reasoning Performance Improves While Error Rates Decrease

Technological improvements in GPT-6.1 Terra are not limited to speed and cost; the model also exhibits a substantial leap in accuracy and reliability. In the Agents’ Last Exam test, which evaluates professional tasks requiring medium-level reasoning, Terra scored 15 percentage points higher than GPT-5.6 Terra. This improvement is particularly noteworthy because the model achieves these results while operating at roughly one-third of the operational cost of its predecessor. When compared to the flagship GPT-6 Astra, Terra maintains similar performance levels despite having a much lower price point per token.

Error reduction is another area where OpenAI has made significant strides. During internal testing with error-prone prompts, the failure rate for low-reasoning tasks plummeted from 53.6 percent in the previous generation to just 16.7 percent in GPT-6.1 Terra. At the highest reasoning settings, the error rate reached an impressive low of 4.3 percent. This increased precision is also evident in software development benchmarks. In the Terminal-Bench 4.0 test, which measures the ability to solve software issues via a command-line interface, the model successfully resolved 50.3 percent of tasks. This is more than double the 21.5 percent success rate recorded by GPT-5.6 Terra.

Security and Accessibility Features Support Enterprise Integration

GPT-6.1 Terra is designed with a massive 1.05 million token context window, allowing it to process vast amounts of information in a single session. This is particularly useful for analyzing long technical documents or complex codebases. Developers can access the model via the OpenAI API under the name “gpt-6.1-terra,” and it fully supports the computer use tool through the Responses API. While the model is highly capable, OpenAI still recommends GPT-6.1 Sol and GPT-6 Astra for the most intricate and complex coding challenges that require the highest possible reasoning depth.

From a security perspective, OpenAI ensures that GPT-6.1 Terra adheres to strict safety protocols. The model has met the “Critical” threshold for cybersecurity within the company’s Preparedness Framework. Furthermore, it inherits all the safety guardrails developed for GPT-6 Astra, including strong resistance to malicious instructions that attempt to divert the AI from its assigned task. These security measures are essential for enterprise clients who require reliable and safe AI agents to operate within their internal networks and handle sensitive company data during the automation of daily workflows.

What tasks in your daily digital life would you like to hand over to an AI agent like GPT-6.1 Terra? Do you think the reduction in error rates is enough to trust AI with your computer? Share your thoughts and expectations in the comments section below.