OpenAI launched GPT-6.1 Terra on October 26, 2026, positioning the new model as a high-speed, cost-efficient “workhorse” designed specifically for AI agent workflows. The release signals a shift in the company’s hierarchy, moving away from a primary focus on raw reasoning toward the practical execution of tasks like coding, form filling, and terminal-based automation.
The new model is reportedly 3x faster at completing multi-step tasks compared to the flagship GPT-6 Astra and the previous GPT-6.1 Sol. To support high-volume enterprise automation, OpenAI has reduced pricing by 50% compared to its predecessors. Standard API pricing for GPT-6.1 Terra is set at $1.00 per 1 million input tokens and $5.00 per 1 million output tokens.
The model has reached general availability in GitHub Copilot, where it is being utilized for agentic coding and terminal work. It is also available via ChatGPT Work and Codex. In terms of technical capacity, Terra features a massive 1.05-million-token context window and a 128,000-token output limit, allowing it to process and generate significantly more data in a single pass than earlier mid-range models.
Performance and Agentic Capability
OpenAI’s latest internal benchmarks, specifically the “Agents’ Last Exam,” show GPT-6.1 Terra scoring 60.0%. This represents a 15-point lead over the previous GPT-5.6 Terra. Notably, this score matches the more expensive GPT-6 Astra in specific task categories, although OpenAI documentation indicates that Terra remains slightly less capable than Astra when handling deep research or highly complex reasoning.
Beyond speed and cost, OpenAI has categorized Terra as meeting the “Critical” threshold for cybersecurity under its 2025 Preparedness Framework. This classification means the model possesses the technical capability to identify and exploit previously unknown software flaws, a level of power that OpenAI states is managed through integrated safety safeguards.
The introduction of Terra clarifies the 2026 model hierarchy, sitting alongside GPT-5.6 Terra and GPT-6.1 Sol. While Astra remains the primary choice for research and creative synthesis, Terra is optimized for “computer-use” capabilities, where the priority is rapid interaction with software interfaces and high-frequency API calls.
According to 9to5Mac, the 50% price reduction is expected to lower the barrier for businesses looking to deploy autonomous agents that run continuously in the background of corporate environments.
