📊 Browse AI model rankings on AI Rank

📱 Get the AI Rank app

GPT-6.1 Terra (also searched as ChatGPT 6.1 Terra) is OpenAI’s fast, efficient model for visual reasoning, precise computer use, and clearer writing. OpenAI announced it on openai.com/index/introducing-gpt-6-1-terra. API model id: gpt-6.1-terra. For earlier GPT-6 family context on this site, see GPT-6. AI Rank also published a same-week News item under the slug openai-gpt-6-1-terra-1026.

What OpenAI says GPT-6.1 Terra is for

From the launch post (checked October 28, 2026):

  • Combines detailed visual reasoning and computer use with clearer, more natural writing.
  • Framed as a step up from GPT-5.6 Terra for day-to-day work: drafting proposals, reading technical drawings, filling forms, editing videos.
  • Positioned on the cost–intelligence frontier: standard API prices are stated as 50% lower than GPT-5.6 Terra and GPT-6.1 Sol.
  • In ChatGPT Work and Codex, the same plan allowance is said to cover more tasks than GPT-5.6 Terra and substantially more than GPT-6.1 Sol or GPT-6 Astra.

Developer docs at developers.openai.com/api/docs/models/gpt-6.1-terra add:

  • Input: text, image · Output: text (audio/video unsupported)
  • Context window 1,050,000 · max input 922,000 · max output 128,000
  • Knowledge cutoff Jun 01, 2026
  • reasoning.effort: none, low, medium (default), high, xhigh, max
  • Tool calling via the Responses API; supported tools include web_search, file_search, computer_use, code_interpreter, MCP, and others listed on that page

API pricing (standard)

Numbers below match both the launch post and the model docs page when checked October 28, 2026. Prefer the live docs if they diverge later.

  • Input: $1 / 1M tokens
  • Cached input: $0.05 / 1M tokens
  • Output: $5 / 1M tokens

Model docs also note cache-write rates (30-minute retention $1.25 / 1M; 12-hour $2.50 / 1M before long-context adjustments), 2× input/cache and 1.5× output when prompts exceed 272K input tokens, Batch/Flex at 50% of Standard, regional processing +10% where available, and Fast mode at 2× Standard.

How OpenAI compares Terra to Sol, Astra, and GPT-5.6 Terra

All figures below are OpenAI’s published claims from the launch article—not AI Rank re-runs.

  • Agents’ Last Exam (computer use): at medium effort, GPT-6.1 Terra scores 15 percentage points higher than GPT-5.6 Terra at about one-third the cost; described as on par with GPT-6 Astra at about one-tenth its price per token. At max effort, Terra is framed as over 3× as token-efficient as Claude Opus 5.5 and about one-fourth its price per token (~9× lower cost per task), and over 3× faster than GPT-6 Astra and GPT-6.1 Sol.
  • MMMU-Pro (visual reasoning): outperforms GPT-5.6 Terra at every reasoning setting; nearly matches GPT-6 Astra and GPT-6.1 Sol at lower cost.
  • OSWorld 2.1: at medium effort, surpasses GPT-5.6 Terra’s best score at a fraction of the cost per task.
  • Internal factuality / error-inducing prompts: low-effort error rate 53.6% → 16.7% vs GPT-5.6 Terra; max-effort 4.3% (18-point drop), nearly matching Astra and Sol.
  • Terminal-Bench 4.0 (coding agents): max effort 50.3% vs GPT-5.6 Terra 21.5%. OpenAI still points complex coding to GPT-6.1 Sol and GPT-6 Astra.
  • Safety: meets the Critical cybersecurity threshold under OpenAI’s Preparedness Framework; keeps GPT-6 Astra protections. System card linked from the launch post.

Availability

Per the launch post:

  • ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users
  • API id gpt-6.1-terra

Snapshot alias on the docs page: gpt-6.1-terra.

→ Rankings: compare related models on the AI Rank model rankings

📱 Track them in the AI Rank app

FAQ

What is the GPT-6.1 Terra API id?

gpt-6.1-terra, per OpenAI’s launch post and model docs.

How much does GPT-6.1 Terra cost?

Standard API: $1 / M input, $0.05 / M cached input, $5 / M output (checked October 28, 2026). Long-context, Batch/Flex, regional, and Fast multipliers are on the model docs page.

How does GPT-6.1 Terra relate to GPT-6.1 Sol and GPT-6 Astra?

OpenAI prices Terra lower and markets it for high-volume computer-use and visual work. Sol and Astra remain the pointers for the heaviest coding / frontier tasks in the same launch copy.

Is GPT-6.1 Terra in ChatGPT?

Yes for ChatGPT Work (and Codex) for Plus, Pro, Enterprise, and Edu, per the launch post—not a claim about every ChatGPT surface.

Sources

Keep exploring: check live model rankings on AI Rank models.

Or 📱 get the AI Rank app.

Prepared with automated editorial assistance; numbers taken from OpenAI pages opened October 28, 2026 (America/Los_Angeles). We did not re-run Agents’ Last Exam or Terminal-Bench ourselves.