OpenAI has shipped GPT-6.1 Terra, a model built for fast, cheap computer use rather than raw reasoning. According to the launch thread, Terra understands what is on screen, takes action across apps and writes in a natural voice, which makes it suited to workday automation. It is rolling out to all Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and developers can call it through the OpenAI API as gpt-6.1-terra. OpenAI positions it as a successor to GPT-5.6 Terra rather than a rival to its own frontier models.
The headline numbers are about cost and speed. On Agents' Last Exam, a benchmark for how well models operate professional software, Terra reportedly matches GPT-6 Astra while costing one tenth as much per token and finishing tasks more than three times faster than both Astra and GPT-6.1 Sol. Standard API prices are 50 percent lower than GPT-5.6 Terra, with the launch graphic listing roughly USD 1 per million input tokens and USD 5 per million output tokens. GPT-6.1 Terra also entered LMArena's Agent Arena and Code Arena on release.
For founders the appeal is unit economics: computer-use agents that previously needed frontier-model budgets can now run at high volume. OpenAI still recommends GPT-6.1 Sol or GPT-6 Astra for deep research and complex coding, so the practical pattern is to route routine screen-driven work to Terra and escalate harder jobs. Cheap, fast, dependable computer use is the piece most agent startups have been waiting for, and this release pushes it closer to commodity pricing.