OpenAI launches GPT-6.1 Terra with faster computer-use
OpenAI launched GPT-6.1 Terra on October 26, 2026, expanding its agent-capable model lineup with a focus on speed, efficiency, and real-world task execution. The model combines detailed visual reasoning with precise computer-use capabilities, allowing agents to act across browsers, desktop software, and applications, OpenAI announced.
New Capabilities in Computer Use
Terra's computer-use system represents a refinement in how agentic models interact with digital environments. The model can process visual input to reason about on-screen elements and execute actions—a capability critical for agents handling workflows that span multiple applications and interfaces. OpenAI says Terra delivers "clearer, more natural writing" alongside these technical capabilities, positioning it as suitable for professional work.
The rollout includes access through the OpenAI API as `gpt-6.1-terra`, with parallel deployment across ChatGPT Work and Codex. Plus, Pro, Enterprise, and Edu users gain access, broadening the distribution of agentic capabilities across OpenAI's user base.
Pricing and Developer Economics
OpenAI priced Terra competitively within its model suite: $1 per million input tokens, $5 per million output tokens, and $0.05 per million cached input tokens per the official announcement. The cached-token rate reflects OpenAI's optimization for repeated reasoning over long documents or workflows—a pattern common in agent-driven automation where context reuse reduces computational overhead.
This pricing aligns with OpenAI's broader strategy to lower barriers for agent adoption, particularly for developers building multi-step automation systems where token costs accumulate rapidly.
Market Context
Terra arrives as the AI agent economy continues to mature. Enterprise adoption of agentic systems remains constrained by model cost, reliability, and the complexity of integrating agents into existing workflows. A model optimized for both speed and price addresses two of these friction points. The emphasis on visual reasoning and computer use suggests OpenAI is targeting use cases where agents must interpret dynamic UIs and execute precise actions—data entry, content moderation, testing automation, and customer-service workflows.
The launch also signals OpenAI's confidence in the GPT-6 series as a stable, production-ready foundation for agent builders. Unlike earlier experimental releases, Terra ships with documented access patterns and pricing, signaling availability for immediate integration into developer workflows and commercial products.
Availability
Developers can access Terra immediately via the OpenAI API. End users will encounter the model through ChatGPT Work and Codex, though the extent and timing of rollout across all Codex instances remain to be detailed by OpenAI.