OpenAI Releases Low-Cost Model GPT-6.1 Terra, Cost per Task About One-Ninth That of Claude Opus 5.5

OpenAI's low-cost new model GPT-6.1 Terra makes a late-night splash: cost per task just one-ninth of Claude Opus 5.5, faster and more accurate

2026-10-27 07:26 IT Home - Qingyuan

English text is being prepared.

IT Home, October 27 news: OpenAI announced in the early hours of today (the 27th) the launch of a new model, GPT-6.1 Terra, focused on completing complex computer operation tasks at lower cost, and especially adept at work requiring visual reasoning and operating desktop software. Compared with GPT-5.6 Terra, the new model has also improved substantially in factual accuracy and coding capability.

Terra is aimed mainly at enterprises looking to reduce automation costs, and can be used for tasks such as processing invoices and interpreting technical drawings. OpenAI says that even for highly difficult professional tasks, Terra can complete them at a cost far below that of high-priced models.

On the Agents' Last Exam benchmark, which assesses complex professional work, Terra's results improved markedly. Test items include building three-dimensional models from architectural drawings and logging film shots one by one for use by film restoration teams.

At medium reasoning effort, Terra scores 15 percentage points higher than GPT-5.6 Terra, while the cost of completing tasks is only about one-third of the latter's. Compared with GPT-6 Astra, Terra's test scores are on par, yet its price per Token is only about one-tenth of the latter's.

Terra also has an advantage in speed. OpenAI says its task execution speed is more than 3 times that of GPT-6 Astra and GPT-6.1 Sol. At the highest reasoning effort, the cost per task is also only about one-ninth of Claude Opus 5.5's.

Desktop software automation is another area where Terra has made progress. OSWorld 2.1 tests a model's ability to use desktop applications through multi-step operation tasks. At medium reasoning effort, Terra not only surpassed GPT-5.6 Terra's previous best result, but also completed each task at a markedly lower cost.

On the visual reasoning benchmark MMMU-Pro, Terra's results at all reasoning efforts exceed those of GPT-5.6 Terra. Despite the lower cost, its performance remains close to GPT-6 Astra and GPT-6.1 Sol.

OpenAI also focused on reducing the model's factual errors to lessen the impact of incorrect answers on automated workflows. The company used an internal benchmark suite with specially designed prompts to probe weak points in the model's answers.

The results show that at low reasoning effort, Terra's error rate is 16.7%, far below GPT-5.6 Terra's 53.6%, a drop of 36.9 percentage points.

After raising reasoning effort to the highest level, Terra's error rate falls further to 4.3%, 18 percentage points lower than the previous generation and close to the level of GPT-6 Astra and GPT-6.1 Sol.

Coding capability has also improved significantly. Terminal-Bench 4.0 tests AI coding agents' ability to solve problems through command-line tasks, such as fixing layout shifts that occur when a web page loads.

At the highest reasoning effort, Terra successfully completed 50.3% of test tasks, while GPT-5.6 Terra's success rate was only 21.5%, meaning the new model's success rate is more than double that of its predecessor. For the most difficult coding tasks, OpenAI still recommends choosing GPT-6.1 Sol or GPT-6 Astra.

GPT-6.1 Terra opens to Plus, Pro, Enterprise, and Edu users on October 27, and can be used in ChatGPT Work and Codex. Developers can also call it through the OpenAI API, with the model identifier gpt-6.1-terra.

For standard API pricing, it charges $1 per million input Tokens (IT Home note: about 6.7 yuan at current exchange rates), $0.05 per million cached input Tokens (about 0.33 yuan at current exchange rates), and $5 per million output Tokens (about 33.4 yuan at current exchange rates). OpenAI says the new model's price is 50% lower than GPT-5.6 Terra and GPT-6.1 Sol.

On safety, Terra follows the safeguards of GPT-6 Astra. OpenAI says the new model has enhanced resistance to prompt injection attacks and can also explain its own limitations to users more clearly. Evaluated under OpenAI's Preparedness Framework, Terra's cybersecurity capability has reached the "critical" threshold.

The lower price is expected to make it easier for developers and enterprises to adopt complex computer automation technology. For the most difficult coding tasks, GPT-6.1 Sol and GPT-6 Astra still have stronger processing capability.