OpenAI's GPT-6.1 Terra makes ChatGPT Work faster and stretches your Plus plan further

What happened

On October 26, OpenAI introduced GPT-6.1 Terra, which it calls “a fast, efficient model that combines detailed visual reasoning and precise computer use alongside clearer, more natural writing.” OpenAI pitches it at everyday work tasks like drafting proposals, interpreting technical drawings, filling out forms and editing videos (OpenAI).

Where you can use it: Terra is rolling out in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu plans. You pick it from the model list, and your plan’s overall usage limits don’t change (ChatGPT release notes). It isn’t available in regular Chat, and Free and Go users don’t get it. In the desktop app or Work on the web it sits under Advanced in the model list (ChatGPT Learn). OpenAI’s release notes also say some Plus and Pro users may see Terra at Medium become the default in Work’s “Power” preset (ChatGPT release notes).

What it costs: OpenAI says that in ChatGPT Work and Codex “the same usage allowance covers more tasks” with Terra than with GPT-5.6 Terra, and “substantially more” than with GPT-6.1 Sol or GPT-6 Astra. For developers, API prices are 50% lower than GPT-5.6 Terra and GPT-6.1 Sol: $1 per million input tokens and $5 per million output tokens (OpenAI).

How good it is, by OpenAI’s numbers: on the Agents’ Last Exam benchmark, Terra scored 15 percentage points higher than GPT-5.6 Terra at medium effort for about a third of the cost. OpenAI also says it performs on par with GPT-6 Astra at a tenth of the price per token, and finishes tasks over three times faster than Astra and GPT-6.1 Sol (BleepingComputer). These are OpenAI’s own figures; we haven’t tested it yet.

It’s also generally available in GitHub Copilot for paid plans, with a gradual rollout (GitHub Changelog).

Our take

This is a price story dressed up as a model story. The biggest complaint about agents like ChatGPT Work is that they’re slow and burn through your limits. A model that’s quicker and lets the same $20 Plus plan do more tasks matters more to most people than another benchmark win. If you’ve been rationing Work, try switching the model to Terra for routine jobs.

Pick the model by the job, not by the name. OpenAI now has Astra, Sol, Terra and Luna, which is confusing. A simple rule: use Terra for screen and form work, sorting photos and documents, and first drafts; keep the bigger model for hard thinking like strategy, a tricky contract or messy numbers. If a Terra result looks shallow, rerun just that task on the bigger one.

For founders and teams: this is the kind of model that makes boring back-office automation cheaper, such as reading invoices and typing them into your accounting tool, which is literally OpenAI’s own example (OpenAI). Before you hand it anything that clicks “submit” or “pay”, watch it do the task a few times and keep a human approval step. Business and Enterprise admins also control which models staff can use, so check your workspace settings (ChatGPT Learn).

For everyday users: if you’re on Free or Go, nothing changes for you today. If you’re deciding whether to upgrade, our Is ChatGPT Plus worth it? guide is updated with Terra. If you’re comparing agents, Claude and Microsoft offer their own work agents too; see Claude Cowork vs ChatGPT Work.

Disclosure: AgentStackWatch is run with help from AI assistants, and we test ChatGPT, Claude, Gemini and Grok Bot. Nobody named here paid for or reviewed this post.