OpenAI Bets GPT-6.1 Terra Can Make AI Agents Cheaper—and Faster
OpenAI Bets GPT-6.1 Terra Can Make AI Agents Cheaper—and Faster
OpenAI on Tuesday introduced GPT-6.1 Terra, a new model aimed squarely at the unglamorous but costly work of the modern office: reading invoices, filling forms, interpreting technical drawings, editing video and moving between desktop applications. The company said Terra combines “detailed visual reasoning and precise computer use” with more natural writing.
The release is the latest move in OpenAI’s push to turn AI from a conversational assistant into an agent that can navigate a screen and complete multistep tasks. Terra is available immediately to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, as well as through the API under the gpt-6.1-terra name.
OpenAI’s pitch rests as much on price as performance. It says standard API pricing is 50% below GPT-5.6 Terra and GPT-6.1 Sol, while its own testing found the model scored 15 percentage points higher than GPT-5.6 Terra on Agents’ Last Exam at roughly a third of the cost. The company also reported lower error rates and stronger resistance to prompt-injection attempts, though it said its more capable Sol and Astra models remain preferable for the hardest coding work.
That framing was echoed in coverage describing the model as “a fast, efficient model” built for visual reasoning, computer use and workplace communication. But OpenAI’s executives pushed the promise further. Sam Altman called Terra the start of “this new era of computer use.”
Greg Brockman made the sharper personal claim: computer use with Terra “has crossed a qualitative threshold where it feels faster at using a computer than what i can do unassisted.” That is the real test now facing Terra—not merely whether it posts strong benchmark numbers, but whether professionals will trust it to take the mouse, read the screen and get the work done.
Write a comment