OpenAI has launched GPT-6.1 Terra, a new model built for computer use, visual reasoning, coding, and the kind of routine digital work that fills an average office day — and it costs half as much as the model it replaces. The company says the model delivers results comparable to its frontier GPT-6 Astra at about one-tenth the price per token, while completing tasks more than three times faster than both GPT-6 Astra and GPT-6.1 Sol.
The launch, announced October 26, is the latest move in OpenAI's push to make agentic AI affordable enough for high-volume, everyday use rather than flagship-only workloads. It comes as competitors slash prices and enterprises scrutinize every token on their invoices. For continued coverage of new model releases, follow our latest AI developments.
What GPT-6.1 Terra is built for
"We're introducing GPT-6.1 Terra, a fast, efficient model that combines detailed visual reasoning and precise computer use alongside clearer, more natural writing," OpenAI said in its announcement, as quoted by BleepingComputer. "A significant step up from GPT-5.6 Terra, it excels at the communication and computer-based tasks that fill the workday, from drafting proposals and interpreting technical drawings to filling out forms and editing videos."
In practice, that positions Terra as OpenAI's workhorse model for agentic computing: the model an enterprise reaches for when it needs to operate software interfaces, process documents, and automate repetitive desktop tasks at scale — workloads where paying frontier-model prices would make the economics unworkable.
The pricing: $1 per million input tokens
Terra's API pricing is set at $1 per million input tokens and $5 per million output tokens, which OpenAI says is 50% lower than both GPT-5.6 Terra and GPT-6.1 Sol. The cut continues a pattern: in July, OpenAI slashed API prices on its GPT-5.6 Luna and Terra models by up to 80%, a move that kicked off the current price war among frontier labs.
The performance claims are where OpenAI is making its case. On the Agents' Last Exam benchmark, Terra scored 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort while costing roughly one-third as much. OpenAI says it performs on par with GPT-6 Astra at one-tenth of Astra's per-token price, and completes tasks more than three times faster than either GPT-6 Astra or GPT-6.1 Sol.
Fewer errors, better coding scores
Hallucination and error rates — a persistent pain point for models doing unsupervised computer work — improved substantially in OpenAI's internal testing. At low reasoning effort, the error rate fell from 53.6% with GPT-5.6 Terra to 16.7% with GPT-6.1 Terra. At maximum effort, the error rate dropped by 18 percentage points to 4.3%, which OpenAI says nearly matches the rates of GPT-6 Astra and GPT-6.1 Sol.
Coding scores improved as well. Terra notched 50.3% on Terminal-Bench 4.0 at maximum reasoning effort, more than double GPT-5.6 Terra's 21.5%. Still, OpenAI continues to recommend GPT-6.1 Sol or GPT-6 Astra for the most complex coding tasks, positioning Terra as the efficiency play rather than the ceiling.
Where it's available
GPT-6.1 Terra is rolling out now to ChatGPT Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex, and is available to developers through the API under the model name gpt-6.1-terra.
The safety caveat hanging over the launch
The price-performance story is not the only one being told about Terra. Related coverage from BleepingComputer notes that OpenAI's testing shows the model is safer overall than its predecessors, but flags more destructive actions in safety evaluations — a reminder that faster, cheaper, more autonomous models multiply both the productive and the risky things an agent can do.
That tension is becoming the defining question of the agentic AI era. A model that can fill out forms, edit videos, and operate software three times faster is also a model whose mistakes — and whose susceptibility to misuse — arrive three times faster. OpenAI's benchmark gains will be scrutinized against that backdrop, particularly as enterprises move from pilots to production deployments where an agent's error rate translates directly into operational risk.
What it means for the price war
The bigger picture is OpenAI's accelerating commoditization of its own product line. With GPT-6 Astra anchoring the frontier, GPT-6.1 Sol covering complex coding, and Terra now undercutting both on price, OpenAI is building the kind of tiered lineup that enterprise buyers have long asked for — and that pressures rivals like Anthropic and Google to keep pace on pricing. In July, OpenAI's 80% cuts to GPT-5.6 pricing were widely read as a shot at Anthropic's enterprise base. Terra's launch extends that strategy deeper into the agent-workload market, where token volume — and therefore token price — determines who can afford to deploy AI at scale.
For buyers, the message is straightforward: the cost of competent AI labor is falling fast, and the gap between frontier performance and affordable performance is closing with every release.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
