OpenAI Releases GPT-6.1 Terra with Advanced Visual Reasoning and Computer Use
OpenAI introduced GPT-6.1 Terra, a mid-tier model optimized for visual reasoning, desktop computer use, and natural communication. Priced at $1 per million input tokens and $5 per million output tokens, it delivers significant cost and speed improvements over its predecessor, GPT-5.6 Terra. Positioned similarly to Anthropic's Sonnet models, GPT-6.1 Terra fills a crucial niche for enterprises seeking affordable, automated execution of repetitive desktop tasks. Its lower price point and high processing speed make it ideal for cost-sensitive workflows like form-filling, video editing, and technical drawing interpretation. GPT-6.1 Terra is priced at nearly half the cost of GPT-6.1 Sol while offering much faster execution on vision and computer-use workloads. Technical commentators note that while it trades off broader general performance, it excels notably in communication and technical reasoning tasks.
## BACKGROUND
Computer use refers to an AI model's ability to directly operate software interfaces, web browsers, and desktop applications like a human user. Visual reasoning involves an AI's capability to mentally interpret, analyze, and process visual inputs such as diagrams, technical drawings, and user interfaces to solve complex problems.