OpenAI introduced GPT-6.1 Terra on Thursday, a new artificial intelligence system built to handle daily workplace duties with greater speed, reduced errors, and lower expenses than its predecessor.
According to the firm, this system enhances writing, programming, visual comprehension, and computer operations, helping users generate documents, interpret technical drawings, complete forms, and modify videos. Standard API rates for GPT-6.1 Terra are also 50% cheaper compared to GPT-5.6 Terra and GPT-6.1 Sol.
Faster Performance and Better Computer Use
OpenAI states that GPT-6.1 Terra executes complex operations spanning multiple programs, which includes turning architectural diagrams into 3D models and analyzing videos frame by frame. During the Agents' Last Exam test, the system achieved a score 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort while running at roughly one-third of the cost per task. It also matched the performance of GPT-6 Astra at approximately one-tenth of its price per token. OpenAI claims the model finishes assignments more than three times faster than GPT-6 Astra and GPT-6.1 Sol. The system also displayed visual comprehension enhancements, nearly equalling OpenAI's more powerful models in the MMMU-Pro test while running at reduced costs.
GPT-6.1 Terra: delivering fast and capable computer use at a lower price.
GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it. pic.twitter.com/8JTN1yonmV
— OpenAI (@OpenAI) October 26, 2026
Major Improvements in Accuracy
OpenAI reports a substantial decrease in factual errors made by GPT-6.1 Terra. During internal evaluations featuring difficult questions, the error rate of the model dropped significantly against GPT-5.6 Terra. At low reasoning effort, the factual error rate declined by nearly 37 percentage points. At maximum effort, it dropped by 18 percentage points to 4.3%, nearing the accuracy of GPT-6 Astra and GPT-6.1 Sol within the same assessment. These findings stem from internal testing by OpenAI and do not guarantee identical accuracy in every scenario.
Better Coding Performance
GPT-6.1 Terra delivers major upgrades in software creation as well. In Terminal-Bench 4.0, an evaluation tracking how effectively AI agents finish programming tasks, the model resolved 50.3% of problems at maximum reasoning effort. This figure exceeds double the performance of GPT-5.6 Terra. The system also demonstrated strong results in scientific programming tests, beating GPT-6.1 Sol in one evaluation while remaining cheaper. Nevertheless, OpenAI notes that GPT-6.1 Sol and GPT-6 Astra continue to be superior options for the most intricate programming tasks.
Improved Safety and Reliability
OpenAI shares that GPT-6.1 Terra excels at obeying directions, identifying boundaries, and stating when it cannot finish a job. The company also highlights enhanced defense against malicious instructions intended to manipulate AI platforms. The model satisfies OpenAI's Critical cybersecurity capability threshold under its Preparedness Framework and incorporates safety safeguards built for GPT-6 Astra.
Pricing and Availability
GPT-6.1 Terra is accessible to ChatGPT Plus, Pro, Enterprise, and Edu users through ChatGPT Work and Codex. OpenAI mentions that subscribers can finish more jobs within their current usage limits compared to GPT-5.6 Terra, GPT-6.1 Sol, and GPT-6 Astra. Software builders can reach the system through the OpenAI API by using the identifier gpt-6.1-terra. OpenAI states that the reduced pricing will render the system more useful for companies needing to automate numerous repetitive duties, such as reading invoices, inputting data into accounting platforms, and handling documents.
