OpenAI launches GPT-6.1 Terra for faster computer-based tasks

A step up in efficiency

OpenAI has officially introduced GPT-6.1 Terra, a new artificial intelligence model built to manage common work tasks with greater speed. The company describes the new model as a "significant step up" from the previous GPT-5.6 Terra version. It aims to provide a more effective balance between performance, cost, and how quickly it completes requests.

The model is designed to handle a wide variety of professional duties that require interacting with different types of data. This includes drafting formal proposals, reading complex technical drawings, and filling out various forms. It is also capable of performing more creative and technical edits, such as editing videos or creating 3D models from architectural plans.

To achieve these goals, OpenAI integrated visual understanding with the ability to use computer applications. This allows the AI to see what is happening on a screen and take actions across different software programs. The company intends for the model to be a versatile tool for a user's entire workday rather than just a single-purpose chatbot.

Benchmarks and accuracy

The performance gains of GPT-6.1 Terra are backed by several internal and external tests. As Digit reported, the model outperformed GPT-5.6 Terra by 15 percentage points on the Agents’ Last Exam benchmark when using medium reasoning effort. Notably, it managed to reach this score while only costing one-third of the previous model's price.

OpenAI also focused on reducing the frequency of factual errors in the new version. On their internal benchmarks, the error rate for tasks requiring low reasoning effort dropped from 53.6 per cent in the older model to just 16.7 per cent. When the tasks required maximum reasoning effort, the error rate fell even further to 4.3 per cent.

Speed is another primary metric where the new model excels over other offerings in the lineup. OpenAI stated that GPT-6.1 Terra completes tasks more than three times faster than both GPT-6 Astra and GPT-6.1 Sol. These improvements make it a more viable option for high-volume tasks where time and accuracy are both critical factors for the user.

Coding and software development

The new model also brings notable improvements to software development and coding capabilities. In the Terminal-Bench 4.0 coding benchmark, GPT-6.1 Terra solved 50.3 per cent of tasks at maximum reasoning effort. This represents more than double the 21.5 per cent score that was published for the older GPT-5.6 Terra model.

While the improvements in coding are significant, OpenAI clarified that this model is not intended to replace every specialized tool. The company noted that GPT-6.1 Sol and GPT-6 Astra remain the better choices for the most complex coding tasks. This suggests that Terra is positioned as a balanced generalist rather than a niche specialist for heavy engineering.

By providing a more capable model for software tasks at a lower price point, OpenAI is making it easier for developers to integrate AI into their workflows. The model can handle a variety of coding requests while maintaining a lower cost structure. This helps teams scale their development processes without incurring the high costs associated with more complex models.

Pricing and availability

OpenAI is making GPT-6.1 Terra available to several different types of users. It can be accessed by Plus, Pro, Enterprise, and Edu users through the ChatGPT Work and Codex platforms. Developers who want to build their own applications can also access the model via the OpenAI API under the specific name gpt-6.1-terra.

The pricing for the API is structured to be highly competitive for high-volume users. It costs $1 for every million input tokens, $0.05 for every million cached input tokens, and $5 for every million output tokens. OpenAI highlighted that this standard API pricing is 50 per cent lower than the pricing for both GPT-5.6 Terra and GPT-6.1 Sol.

By lowering the cost of these tokens, OpenAI is making it more affordable for companies to run large-scale operations. These savings are intended to help users take on more of their workday tasks using AI. The combination of lower costs and faster speeds makes it a practical choice for businesses looking to automate repetitive computer-based work.