OpenAI introduced GPT-6.1 Terra on October 26, 2026, to help speed up digital workflows. This model interacts with apps to click and type like a human. Sam Altman said, "6.1 Terra is here! It is good at a lot of things, but we are especially excited for this new era."
OpenAI has just introduced its latest artificial intelligence model, GPT-6.1 Terra. Launched on October 26, 2026, this new AI is not just another chatbot that answers your questions. Instead, it acts as an active digital assistant that can actually look at your computer screen, click on things, and complete your everyday work.
Celebrating the launch on X (formerly Twitter), OpenAI CEO Sam Altman posted, "6.1 Terra is here! It is good at a lot of things, but we are especially excited for this new era of computer use."
How GPT-6.1 Terra works on your computer
Unlike older AI models that only generate text, GPT-6.1 Terra is built to interact directly with your computer software. It can visually understand what is showing on your screen and use the interface just like a normal user would. It can click, type, and switch between different applications to finish a multi-step task.
For example, businesses can use it to automatically copy invoice details into accounting software, fill out long online forms, prepare business proposals, or even turn architectural drawings into 3D models. The main goal is to let the AI handle boring, repetitive digital tasks without human help.
Faster performance and lower costs
The new model is a major upgrade over the older GPT-5.6 Terra. OpenAI claims that GPT-6.1 Terra completes tasks roughly three times faster on average than its heavier, more advanced models like GPT-6.1 Sol and GPT-6 Astra. It is also much cheaper for developers who want to build new applications using it.
OpenAI has slashed standard API prices by 50% compared to the older version. It now costs just $1 per million input tokens and $5 per million output tokens, making it a highly pocket-friendly option for businesses running automated workflows at scale.
Big improvements in accuracy and coding
OpenAI has also made sure that the new Terra is much smarter and makes fewer mistakes. In the company's internal tests, the AI's error rate for simple tasks dropped massively from 53.6% in the older model to just 16.7% in the new one. When pushed to its maximum reasoning effort, the error rate fell to a tiny 4.3%.
It has also shown major progress in coding, scoring more than double what the older version managed on complex terminal benchmarks. Because of this boost, GPT-6.1 Terra is already rolling out to programmers through platforms like GitHub Copilot.
Where does Terra fit in OpenAI's lineup?
OpenAI currently offers a wide family of AI models, which includes Luna, Terra, Sol, and Astra. While Luna is a lightweight option and Sol and Astra are designed for highly complex, heavy-duty work, Terra sits right in the middle as the perfect mid-range model.
Following last month's updates to Sol and Luna, this new release finally brings Terra up to the GPT-6.1 generation. Regular users and businesses can start experiencing GPT-6.1 Terra right away, as it is now available through ChatGPT Plus, Pro, Enterprise, and Edu subscriptions in ChatGPT Work and Codex, as well as OpenAI's API and Amazon Bedrock.