--- title: "OpenAI Unveils GPT-6.1 Terra: Three Times Cheaper and Noticeably Smarter Than Its Predecessor" description: "On October 27, OpenAI introduced GPT-6.1 Terra — a model for everyday tasks that outperforms GPT-6.6 Terra in benchmarks and costs three times less. Available in ChatGPT, Codex, and via API." date: 2026-10-27T11:28:39.984Z lang: en url: https://xab.info/en/posts/openai-unveils-gpt-6-1-terra tags: [openai, gpt-6-1-terra, ai-model, artificial-intelligence, chatgpt] author: "Alexander Buran" publisher: "XAB.info" --- # OpenAI Unveils GPT-6.1 Terra: Three Times Cheaper and Noticeably Smarter Than Its Predecessor ![Transparent digital monitor displaying the OpenAI GPT-6.1 Terra logo and a globe on an office desk](https://xab.info/media/2026/10/27/openai-gpt-6-1-terra/openai-gpt-6-1-terra-1.webp) ## Key Points - OpenAI released the GPT-6.1 Terra model on October 27, 2026, for visual analysis, computer control, and text generation. - In the Agents' Last Exam test, the model scores 15 percentage points higher than GPT-5.6 Terra, with task execution costs roughly three times lower. - In Terminal-Bench 4.0, GPT-6.1 Terra solves 50.3% of tasks compared to 21.5% for its predecessor. - The error rate at low reasoning effort dropped from 53.6% to 16.7%, and down to 4.3% at maximum settings. - API prices are reduced by 50%; the model is available in ChatGPT and Codex for Plus, Pro, Enterprise, and Edu tiers, and via API. On October 27, 2026, OpenAI released a new artificial intelligence model, GPT-6.1 Terra. According to the company, the system combines visual information analysis, computer control, and natural text generation, while offering a high level of intelligence at a reduced price. The release is positioned as a tool for everyday working tasks: from preparing business proposals and reading incoming invoices to interpreting technical drawings, filling out forms, and editing video. ### What GPT-6.1 Terra Can Do The new model is focused on performing applied tasks regularly faced by office and professional users. In its list of capabilities, OpenAI highlights working with documents and invoices, interpreting blueprints, filling out electronic forms, and even video editing. The company also emphasizes the model's high efficiency when working with interfaces — meaning scenarios where the AI must "see" the screen and act within it rather than just generate text. ### Benchmarks: Intelligence Boost with Lower Cost In the specialized Agents' Last Exam test, the model scores 15 percentage points higher than GPT-5.6 Terra, while the cost of completing tasks, according to OpenAI, is roughly three times lower. In the Terminal-Bench 4.0 command-line programming test, the system successfully solves 50.3 percent of tasks, more than double its predecessor's result (21.5 percent). ### Combating Hallucinations Developers directed significant resources toward reducing the frequency of model errors. In internal tests with adversarial prompts at low reasoning effort, GPT-6.1 Terra's error rate dropped from 53.6 percent in GPT-5.6 Terra to 16.7 percent. At maximum reasoning settings, the error rate is just 4.3 percent. ### Safety and Abuse Protection Developers have preserved and improved defense mechanisms against malicious instructions and cyber threats. According to OpenAI, GPT-6.1 Terra honestly notifies the user about task execution limitations and adheres better to set behavioral boundaries, which is crucial for corporate use cases. ### Pricing and Availability Standard API usage pricing has been reduced by 50 percent compared to GPT-5.6 Terra. Starting October 27, GPT-6.1 Terra is available to Plus, Pro, Enterprise, and Edu users in ChatGPT and Codex, as well as to developers via the API. ## Fact-Check & Sources - [Втрое дешевле и умнее: OpenAI выпустила новую модель GPT-6.1 Terra](https://www.rbc.ua/ukr/news/vtrichi-deshevsha-i-rozumnisha-openai-vipustila-1793095419.html) - Primary source: release description, benchmarks (Agents' Last Exam +15 pp, Terminal-Bench 4.0 50.3% vs 21.5%), reduction of hallucination errors (53.6%→16.7%, max 4.3%), API prices -50%, availability from Oct 27, 2026. - [OpenAI выпустила новую флагманскую ИИ-модель GPT‑6 Astra. Президент компании назвал этот релиз началом ...](https://meduza.io/feature/2026/09/04/openai-vypustila-novuyu-flagmanskuyu-ii-model-gpt-6-astra-prezident-kompanii-nazval-etot-reliz-nachalom-novoy-ery) - Context: release of GPT-6 Astra (Sep 4, 2026), a different model; does not confirm Terra's figures, used for chronology. - [OpenAI выпустила новую ИИ-модель GPT-6 Astra](https://www.interfax.ru/digital/1113271) - Context: release of GPT-6 Astra (Sep 4, 2026) and the internet isolation bypass incident; does not directly relate to Terra's facts. ## FAQ ### When is GPT-6.1 Terra available? Starting October 27, 2026, the model is available in ChatGPT and Codex for Plus, Pro, Enterprise, and Edu tiers, as well as to developers via the API. ### How does GPT-6.1 Terra differ from GPT-5.6 Terra? According to OpenAI, it scores 15 percentage points higher than its predecessor in Agents' Last Exam, solves 50.3% of tasks in Terminal-Bench 4.0 versus 21.5%, has roughly three times lower task execution costs, and features 50% lower API prices. ### How much have the model's errors been reduced? In internal tests with adversarial prompts, the error rate at low reasoning effort dropped from 53.6% to 16.7%, and stands at 4.3% at maximum settings.