On this page
OpenAI has shipped GPT-6.1 Terra, a mid-tier model aimed at visual reasoning, precise computer use, and clearer day-to-day writing. The headline for ChatGPT users is access—not a new Chat model: Terra sits in ChatGPT Work and Codex, and is not available in regular ChatGPT conversations.
Who can use it—and who cannot
Per OpenAI’s announce, GPT-6.1 Terra is available to Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex. The ChatGPT & Codex changelog (October 26, 2026) also lists Business. Enterprise access still depends on workspace availability and admin controls. The Help Center groups GPT-6.1 Terra with GPT-6.1 Sol, GPT-6 Sol, and GPT-6 Luna as Work／Codex models that do not appear in regular Chat.
On desktop, Work／Codex model picks live in Advanced; Work on web and mobile can select Terra as well. The changelog notes Terra does not support UAE inference residency.
Codex CLI and desktop clients should follow the minimum Codex version required in the ChatGPT & Codex changelog for this rollout—keep the CLI／app updated. Usage burn rates on Work／Codex have not been measured on this site—for reference only.
Related: How Codex Usage Limits Work on Each ChatGPT Plan (October 2026) and Uber Rolls Out ChatGPT Work to More Than 30,000 Employees.
API price and context
Developers reach the model as gpt-6.1-terra. Standard text-token prices on the API docs page:
| Metric | Price per 1M tokens |
|---|---|
| Input | $1 |
| Cached input | $0.05 |
| Cache writes (30-minute retention) | $1.25 |
| Output | $5 |
OpenAI’s announce puts those Standard rates at about 50% below GPT-5.6 Terra and GPT-6.1 Sol. Prompts with more than 272K input tokens are billed at 2× input and cache rates and 1.5× output for the full request. Batch and Flex are 50% of Standard; Fast mode is 2× Standard; regional processing can add a 10% premium where available.
Model card details from the API docs: 1,050,000 context window; maximum input about 922,000 tokens; 128,000 max output; input text + image; output text; knowledge cutoff June 1, 2026. Reasoning effort supports none, low, medium (default), high, xhigh, and max. Tool calling for computer use and related tools goes through the Responses API.
Where Terra sits vs Sol and Astra
OpenAI frames GPT-6.1 Terra as a clear step up from GPT-5.6 Terra on computer-use and visual-reasoning benchmarks, with Standard API pricing meant for high-volume automation (invoices, forms, screen workflows). On factuality and coding, the announce reports large gains over GPT-5.6 Terra, but still points complex coding toward GPT-6.1 Sol and GPT-6 Astra.
In ChatGPT Work／Codex, OpenAI says the same plan 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—again OpenAI’s framing, not an independent load test.
Subscription long-context rules for the Sol line are a separate story: GPT-6.1 Sol: No Long Context Fee on Subscription. For coding-agent price／limit comparisons outside OpenAI’s stack, see Claude Code vs Codex: What $20 Gets You and Which to Pick.
What this changes in practice
For teams already living in ChatGPT Work—drafting proposals, reading drawings, filling forms, or driving desktop apps—Terra is the cheaper Work／Codex pick OpenAI wants next to Sol and Astra. The announce leans on Agents’ Last Exam, MMMU-Pro, OSWorld 2.1, and Terminal-Bench numbers as OpenAI’s own measurements; those figures are not re-run here.
For API builders, the $1／$5 Standard lane is the concrete decision point: half the list price of GPT-6.1 Sol on OpenAI’s framing, with a 1.05M context window and image input. Long prompts past 272K input still trigger the 2×／1.5× long-context multipliers on the full request, so high-volume pipelines should budget that step, not only the headline $1／$5 row.
Nothing in the public Help Center or announce moves Terra into regular Chat. Model pickers in Chat stay on the Chat lineup; Work and Codex keep the separate Work／Codex catalog that now includes Terra beside Sol and Luna.