Introducing GPT-6.1 Terra

Why it matters

OpenAI has released GPT-6.1 Terra, a model it positions as fast, efficient, and aimed at everyday work tasks that combine visual understanding with computer use. The company says it improves on GPT-5.6 Terra in factuality, coding, and visual reasoning, while costing less per token than GPT-6.1 Sol and performing on par with GPT-6 Astra at a much lower price. OpenAI highlights gains on benchmarks covering desktop-app automation, command-line software engineering, and scientific research tasks, and reports reduced hallucination rates on its internal error-inducing prompt set. The model is available now in ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users, and through the API for developers. OpenAI also describes safety work, saying the model retains protections from GPT-6 Astra, resists prompt injection, and meets its Critical cybersecurity threshold. The release matters because it pushes capable agentic and computer-use features toward lower cost, high-volume business automation, though OpenAI notes its stronger models remain preferable for the hardest coding work.