OpenAI launches GPT-6.1 Terra with faster performance, 50% lower pricing
What's the story
OpenAI has launched its latest artificial intelligence (AI) model, GPT-6.1 Terra. This is a major upgrade from GPT-5.6 Terra, and it promises to be faster and more efficient. The new model combines advanced visual reasoning capabilities with accurate computer use and improved writing skills. The launch comes as part of OpenAI's ongoing efforts to push the boundaries of AI technology while keeping costs reasonable for developers and users alike.
Cost benefits
More cost-effective than previous models
The new model, Terra, is also more cost-effective than its predecessors.
Its standard API prices are 50% lower than those of GPT-5.6 Terra and GPT-6.1 Sol.
This makes it a practical choice for developers looking to automate high-volume tasks such as processing incoming invoices and entering their details into accounting software.
Task efficiency
Terra outperforms predecessors in professional tasks
Terra has been tested on Agents's Last Exam, a benchmark of professional tasks.
These include converting architectural drawings into a 3D model and reviewing an animated film scene by scene to create a time-coded log for a restoration team.
At medium reasoning effort, GPT-6.1 Terra scores 15% points higher than its predecessor at about one-third the cost.
It even matches the performance of GPT-6 Astra at just one-10th its price per token.
Speed advantage
Much faster than previous versions
Terra is also much faster than its predecessors, completing tasks over three times faster than both GPT-6 Astra and GPT-6.1 Sol.
On MMMU-Pro, a benchmark of visual reasoning, the new model outperforms its predecessor at every reasoning setting.
It even comes close to matching the scores of OpenAI's most advanced frontier models, GPT-6 Astra and GPT-6.1 Sol, at a much lower cost.
App proficiency
Reduced hallucinations and factual errors
In OSWorld 2.1, a test of AI agents' ability to use desktop apps for multistep tasks, GPT-6.1 Terra outperforms its predecessor at medium reasoning effort and a fraction of the cost per task.
The new model has also seen significant improvements in reducing hallucinations and factual errors.
On an internal benchmark of error-inducing prompts, the error rate at low reasoning effort fell from 53.6% with GPT-5.6 Terra to 16.7% with GPT-6.1 Terra, a drop of nearly 37% points.
Coding proficiency
Improvements in software engineering tasks
For software engineering tasks, GPT-6.1 Terra has also shown significant improvement over its predecessor.
In Terminal-Bench 4.0, a test of coding agents' ability to solve problems from the command line, GPT-6.1 Terra solves 50.3% of tasks at maximum reasoning effort, more than double its predecessor's score of 21.5%.
However, for complex coding tasks, the newer model still lags behind GPT-6 Astra and GPT-6.1 Sol in terms of performance.
Security upgrades
Terra also comes with enhanced alignment capabilities
Along with performance improvements, GPT-6.1 Terra also comes with enhanced alignment capabilities.
The new model is more likely to stay within scope and honestly disclose limitations and potential blockers to the user compared to its predecessor.
In prompt-injection tests, it showed strong resistance against malicious instructions designed to divert it from the user's task.
Availability
Now available to users
GPT-6.1 Terra is rolling out to all Plus, Pro, Enterprise, and Edu users through ChatGPT Work and Codex.
Developers can also access the model via the OpenAI API under the name gpt-6.1-terra.
The API pricing is set at $1 per million input tokens, $0.05 per million cached input tokens, and $5 per million output tokens.
Twitter Post
OpenAI says new model is very good at computer use
GPT-6.1 Terra: delivering fast and capable computer use at a lower price.
— OpenAI (@OpenAI) October 26, 2026
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