AI News Feed
Market watch
Samsung Galaxy phones can wirelessly charge other devices with PowerShareSymantec Finds Autonomous 'Skynet' AI Agent Hijacking Exposed AI SystemsTESSAN Magic 65W Charging Station Combines Desk Power and MagSafe ChargingGemini may add bulk chat deletion and a 'Remembered items' managerAmazon cuts CMF Headphone Pro to $55, an all-time lowAndroid Authority poll finds Android loyalty strong as iOS curiosity persists
Samsung Galaxy phones can wirelessly charge other devices with PowerShareSymantec Finds Autonomous 'Skynet' AI Agent Hijacking Exposed AI SystemsTESSAN Magic 65W Charging Station Combines Desk Power and MagSafe ChargingGemini may add bulk chat deletion and a 'Remembered items' managerAmazon cuts CMF Headphone Pro to $55, an all-time lowAndroid Authority poll finds Android loyalty strong as iOS curiosity persists
Large Language Models

OpenAI Releases GPT-6.1 Terra for ChatGPT Work, Codex and API

OpenAI releases GPT-6.1 Terra, a faster, cheaper medium model replacing GPT-5.6 Terra in ChatGPT Work, Codex and API.

AI News Feed · 2026-10-27 03:10

The update follows OpenAI’s GPT-6 rollout last month, when the company introduced Sol and Luna but left Terra out of the release, the report said. OpenAI’s main model classes range, from lowest price and capability to highest, across Luna, Terra, Sol and Astra. GPT-6.1 Terra is positioned as the medium-sized model that balances intelligence and affordability.

OpenAI said pricing for GPT-6.1 Terra is 50% lower than for GPT-5.6 Terra, the model it replaces. The company presented that lower cost as suited to automating computer-based work tasks for less.

OpenAI described GPT-6.1 Terra as a significant step up from GPT-5.6 Terra and said it excels at communication and computer-based tasks that fill the workday, including drafting proposals, interpreting technical drawings, filling out forms and editing videos. The model is available starting today through ChatGPT Work, Codex and the API, according to 9to5Mac.

On OpenAI’s 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 percentage points, the report said. At maximum effort, the error rate fell by 18 percentage points to 4.3%, nearly matching the rates for GPT-6 Astra and GPT-6.1 Sol.