OpenAI Releases GPT-6.1 Terra: Enhanced Computer Operation Capabilities, API Price Cut in Half

OpenAI Releases GPT-6.1 Terra: Better at Operating Computers, API Price Cut in Half

Phoenix Tech Download the App for Exclusive Early Access October 27, 2026 06:16:33 From Beijing

Phoenix Tech News, October 27 (Beijing Time) — OpenAI announced the launch of a new model, GPT-6.1 Terra, with a focus on improving visual understanding and computer operation capabilities. It can be used for tasks such as drafting proposals, interpreting technical drawings, filling out forms, and editing videos. OpenAI said the new model's standard API price is 50% lower than both GPT-5.6 Terra and GPT-6.1 Sol.

OpenAI said GPT-6.1 Terra combines more detailed visual reasoning with more precise computer operation capabilities, and its written expression is clearer and more natural. Developers can use it to automatically handle large numbers of repetitive tasks, such as reading incoming invoices and entering the relevant information into financial software.

In ChatGPT Work and Codex, OpenAI said the same usage allowance lets GPT-6.1 Terra complete more tasks than GPT-5.6 Terra, and significantly more than GPT-6.1 Sol or GPT-6 Astra.

Improved Computer Operation Capabilities

OpenAI evaluated the new model using the Agents' Last Exam benchmark, which covers professional work tasks. Test tasks included converting architectural drawings into 3D models, and reviewing an animated film shot by shot and producing time-stamped notes for the restoration team.

According to OpenAI, at medium reasoning effort, GPT-6.1 Terra scored 15 percentage points higher than GPT-5.6 Terra at about one-third the cost. In this test, the new model performed on par with GPT-6 Astra at one-tenth the price per token.

OpenAI also said GPT-6.1 Terra completes test tasks more than three times as fast as GPT-6 Astra and GPT-6.1 Sol. At the highest reasoning effort, compared with Anthropic's Claude Opus 5.5, the new model's cost per task in this test is about 9 times lower.

On the MMMU-Pro visual reasoning benchmark, OpenAI said GPT-6.1 Terra surpasses GPT-5.6 Terra at all reasoning effort settings, scores close to GPT-6 Astra and GPT-6.1 Sol, and costs far less.

The above performance and cost figures all come from tests published by OpenAI; actual results will vary depending on the task and settings.

Improved Factual Error Rate and Coding Ability

OpenAI said that in an internally designed set of prompts prone to inducing errors, at low reasoning effort, GPT-6.1 Terra's factual error rate dropped from 53.6% for GPT-5.6 Terra to 16.7%. At the highest reasoning effort, the new model's error rate fell to 4.3%, an 18-percentage-point decrease from the previous generation.

In coding, on Terminal-Bench 4.0, which tests AI agents' ability to solve problems via the command line, GPT-6.1 Terra completed **50.3%** of tasks at the highest reasoning effort, more than double the 21.5% published score of GPT-5.6 Terra.

However, OpenAI said that for the most complex coding tasks, GPT-6.1 Sol and GPT-6 Astra remain better choices.

On safety, OpenAI said GPT-6.1 Terra reaches the "critical" capability threshold in the cybersecurity domain under its Preparedness Framework and retains the safeguards developed for GPT-6 Astra. The company's tests also show the new model is more likely than GPT-5.6 Terra to stay within the scope of a task and to honestly explain to users its own limitations and the obstacles it may encounter.

Pricing and Availability

GPT-6.1 Terra is available starting today to Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex. Developers can also call the model through the OpenAI API under the model name gpt-6.1-terra.

Standard API pricing is $1 per million input tokens, $0.05 per million cached input tokens, and $5 per million output tokens. (Author/Yu Lei)

For more first-hand news, please download the Phoenix News app and subscribe to Phoenix Tech. For in-depth reporting, search "Phoenix Tech" on WeChat.