OpenAI has released GPT-6.1 Terra, a reasoning model designed for high-volume visual and computer-use tasks. The model supports text and image inputs, features a context window of 1,050,000 tokens, and has a knowledge cutoff date of June 1, 2026.
The model supports various tools, including computer use, code interpreter, web search, and file search. OpenAI notes that Terra performs tasks over three times faster than the GPT-6 Astra and GPT-6.1 Sol models.
Standard API pricing for the model is set at $1 per million input tokens and $5 per million output tokens. This pricing structure represents a 50 percent reduction compared to the GPT-6.1 Sol model.
In performance testing on the Agents’ Last Exam v1, Terra achieved a 57.6 percent score at medium effort for $0.29 per task. At maximum effort, the model reached a 60.0 percent score at $1.01 per task.
GPT-6.1 Terra is currently available for ChatGPT Plus, Pro, Enterprise, and Edu users within the ChatGPT Work and Codex environments.
