GPT-6.1 Terra: what OpenAI's faster, cheaper model means for ChatGPT, Codex and the API

Short answer: GPT-6.1 Terra is OpenAI's new fast, lower-cost model for computer use, reading images and screens, and clear everyday writing, announced on October 26, 2026. It is available in ChatGPT Work and Codex on paid plans and in the API as gpt-6.1-terra, where OpenAI set its standard price at half that of GPT-6.1 Sol. OpenAI says it handles most routine work at a fraction of the cost of its bigger models, but still points to GPT-6.1 Sol and GPT-6 Astra for the hardest coding tasks.

If you use ChatGPT for agent-style work, write code with Codex or GitHub Copilot, or build on the OpenAI API, Terra changes which model is the sensible default for a lot of jobs. OpenAI's pitch is that much of the workday is communication and computer-based tasks, such as drafting, reading documents and screenshots and filling in forms, and that this work doesn't need its most expensive models.

What OpenAI announced

OpenAI introduced GPT-6.1 Terra on October 26, 2026, describing it as "a fast, efficient model that combines detailed visual reasoning and precise computer use alongside clearer, more natural writing." It is the newest model in OpenAI's Terra line, following GPT-5.6 Terra, and in OpenAI's model list it sits between the top-end GPT-6 Astra and GPT-6.1 Sol and the budget GPT-6 Luna.

The examples OpenAI gives are ordinary office work: drafting proposals, interpreting technical drawings, filling out forms, editing videos, and reading incoming invoices and entering their details into accounting software. OpenAI published a system card addendum the same day.

What's new compared with the last Terra

All of the figures below are OpenAI's own evaluations from its launch post, so treat them as the vendor's claims until independent tests catch up.

• Computer use. On Agents' Last Exam, a benchmark of professional tasks done on a computer, OpenAI says Terra scores 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort, at about a third of the cost, and performs on par with GPT-6 Astra. It also says Terra finishes tasks more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
• Images and screens. On the MMMU-Pro visual reasoning benchmark, OpenAI says Terra beats GPT-5.6 Terra at every reasoning setting and nearly matches GPT-6 Astra and GPT-6.1 Sol.
• Fewer factual errors. On OpenAI's internal set of prompts designed to trigger mistakes, the error rate at low reasoning effort fell from 53.6% with GPT-5.6 Terra to 16.7% with GPT-6.1 Terra.
• Coding. On Terminal-Bench 4.0, which tests command-line coding agents, OpenAI reports 50.3% at maximum effort, more than double GPT-5.6 Terra's 21.5%. OpenAI is clear that "for the most complex coding tasks, GPT-6.1 Sol and GPT-6 Astra remain the better choice."
• Safety. OpenAI says Terra is treated as Critical for cybersecurity capability under its Preparedness Framework and uses the same safeguards as GPT-6.1 Sol and GPT-6 Astra. It also reports strong resistance to prompt injection, the hidden instructions in web pages or files that try to hijack an agent.

Where you can use it

Sources: OpenAI's ChatGPT release notes, Codex changelog, Codex models guide and API model page, plus the GitHub changelog, AWS announcement and Vercel changelog. OpenAI's blog post lists Plus, Pro, Enterprise and Edu; its release notes and Codex changelog also include Business. Access depends on your plan, region and workspace settings.

One regional detail: OpenAI's release notes say Terra does not support UAE inference residency. That depends on where a workspace's processing runs, not where you are, so it mainly matters to organisations that have chosen to keep processing in the UAE.

What it costs

Price is a big part of this launch. OpenAI's API model page lists standard pricing of $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens at launch, which OpenAI says is 50% lower than both GPT-5.6 Terra and GPT-6.1 Sol. The same page notes that very long prompts, Fast mode and regional processing cost more, and that Batch and Flex are cheaper. Check the live OpenAI API pricing page before you budget.

In ChatGPT there is no separate charge. OpenAI's release notes say your plan's overall usage limits are unchanged, while its launch post says the same 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. Some Plus and Pro users may also find Terra at Medium set as a default Power preset in Work, in place of GPT-6.1 Sol at Light. In GitHub Copilot, GitHub says Terra is billed at the provider's list price under usage-based billing.

Terra vs OpenAI's other GPT-6 models

OpenAI's Codex models guide sets out when to use each model in ChatGPT Work and Codex:

In practice, start with Terra for documents, forms, screenshots and routine agent tasks. Move up to Sol or Astra when the job involves a large or messy codebase, a long chain of decisions, or the result has to be right first time.

How it compares with Claude and Gemini

OpenAI compares Terra directly with Anthropic's Claude. It says that at maximum effort Terra is more than three times as token-efficient as Claude Opus 5.5 at a quarter of its price per token, which works out to roughly nine times lower cost per task on Agents' Last Exam. That is OpenAI's own test, not an independent one. Anthropic's models overview currently suggests starting with Claude Opus 5.5 for most workloads, and positions Claude Sonnet 5.5 as its best mix of speed and intelligence and Claude Haiku 5.5 for high-volume, low-latency tasks. Sonnet is the closest match to Terra's role.

On Google's side, the Gemini API models page lists Gemini 4 Argon as Google's flagship reasoning model and Gemini 3.8 Flash as its most intelligent Flash model, aimed at agents and complex enterprise workflows. OpenAI's launch post doesn't compare Terra with Gemini.

Vendor benchmarks rarely settle these choices. The more useful question is which assistant or coding tool you already work in. For side-by-side views of the products, see Claude vs ChatGPT, ChatGPT vs Gemini, Cursor vs GitHub Copilot and GitHub Copilot vs Claude Code.

What it means if you're choosing AI tools

• On ChatGPT Plus or Pro? Try Terra in Work for forms, document review, screenshots and drafting. OpenAI says your allowance should stretch further than with Sol or Astra.
• On ChatGPT Free or Go? These plans aren't part of the rollout, and Terra isn't in the Chat tab, so nothing changes for you yet.
• Writing code? Terra is now an option in Codex and GitHub Copilot for well-scoped tasks and terminal work. For big refactors or tricky bugs, OpenAI itself recommends Sol or Astra. Anthropic's Claude Code and Cursor remain the main alternatives to compare.
• Building on the API? High-volume work such as invoice entry, screenshot QA and form filling is where the lower price matters most. Test it on a sample of your own tasks before switching, because OpenAI's benchmarks may not match your workload.
• In a managed workspace? Enterprise access depends on admin controls, and Copilot Business and Enterprise admins can turn the model on or off in Copilot's model policy.

FAQ

What is GPT-6.1 Terra?

GPT-6.1 Terra is an OpenAI model announced on October 26, 2026. It is built for fast, lower-cost work that involves reading images and screens, using a computer and writing clearly. It follows GPT-5.6 Terra and sits between OpenAI's premium GPT-6.1 Sol and GPT-6 Astra models and the budget GPT-6 Luna.

Is GPT-6.1 Terra available in ChatGPT?

Yes, in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu plans, rolling out gradually. You choose it from the model list. It isn't available in the regular Chat tab, and the Free and Go plans aren't part of the rollout.

How much does GPT-6.1 Terra cost in the API?

At launch, OpenAI lists standard API pricing of $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens, which it says is half the price of GPT-6.1 Sol. Long prompts, Fast mode and regional processing cost more, so check OpenAI's live pricing page.

Is GPT-6.1 Terra good for coding?

It is much better at coding than the previous Terra, according to OpenAI, and it is available in Codex and GitHub Copilot. But OpenAI says GPT-6.1 Sol and GPT-6 Astra remain the better choice for the most complex coding tasks, so Terra suits well-scoped tasks best.

How does GPT-6.1 Terra compare with Claude?

OpenAI says that at maximum effort Terra costs about a ninth as much per task as Claude Opus 5.5 on its Agents' Last Exam benchmark. That is OpenAI's own comparison. In Anthropic's lineup, Claude Sonnet 5.5 is the closest match to Terra's speed-and-cost role, so test both on your own tasks.

Where else can I use GPT-6.1 Terra?

Besides ChatGPT, Codex and the OpenAI API, it was made available on October 26, 2026 in GitHub Copilot, on Amazon Bedrock and through Vercel's AI Gateway.