GPT-6.1 Terra
OpenAI's fast, low-cost GPT-6 model for computer use, visual reasoning and high-volume work
Context window
1.05M
Input / 1M tokens
$1.00
Output / 1M tokens
$5.00
Provider
OpenAI
Cached input $0.05 and cache writes $1.25 (30-minute retention) per 1M. Prompts over 272K input tokens cost $2.00 input / $7.50 output per 1M. Batch and Flex are half price ($0.50/$2.50); Fast mode costs $2.00/$10.00. OpenAI says standard prices are 50% lower than GPT-5.6 Terra ($2.00/$12.00). Max output 128K tokens; maximum input 922K tokens. Knowledge cutoff June 1, 2026. · Data verified 2026-11-03
GPT-6.1 Terra (October 26, 2026) is a fast, efficient model in OpenAI's GPT-6 family, priced between GPT-6.1 Sol and GPT-6 Luna. OpenAI built it for visual reasoning, precise computer use and natural writing, and says it matches GPT-6 Astra on its Agents' Last Exam evaluation at one-tenth the price per token while finishing tasks over 3x faster on average than Astra and GPT-6.1 Sol. It costs $1/$5 per million tokens, accepts text and images, and has a 1.05M-token context window with 128K max output. For in-depth research and complex coding, OpenAI recommends GPT-6.1 Sol or GPT-6 Astra.
How to access it
OpenAI API (model ID gpt-6.1-terra) on the Responses, Chat Completions and Batch endpoints. In ChatGPT it is available to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex.
Strengths
- ✓$1/$5 per 1M tokens, which OpenAI says is 50% below GPT-5.6 Terra's standard price
- ✓Built for computer use and visual reasoning; matches GPT-6 Astra on Agents' Last Exam (OpenAI-reported)
- ✓Over 3x faster on average than GPT-6 Astra and GPT-6.1 Sol (OpenAI-reported)
- ✓1.05M-token context window and 128K max output
- ✓Reasoning effort settings from none to max
Best for developers who...
When to choose it (and when not to)
Reach for GPT-6.1 Terra when...
- →You run computer-use or screen-reading agents at volume
- →You want a GPT-6-generation model at $1/$5 per 1M tokens
- →Latency matters more than peak reasoning depth
Look elsewhere if...
- ✕In-depth research or complex coding (OpenAI recommends GPT-6.1 Sol or GPT-6 Astra)
- ✕Audio or video input (it accepts text and images only)
- ✕Very cheap classification or routing (GPT-6 Luna costs $0.10/$0.50)
How to use it
- ›Set reasoning.effort to none, low, medium, high, xhigh or max; medium is the default
- ›OpenAI's GPT-6 guide covers prompting and migration from GPT-5.6 Terra
Quickstart
Pythonfrom openai import OpenAI
client = OpenAI()
response = client.responses.create(
model="gpt-6.1-terra",
reasoning={"effort": "medium"},
input="List the action items in this meeting summary: ...",
)
print(response.output_text)Responses API. Reasoning effort accepts none, low, medium (default), high, xhigh and max.
API model id: gpt-6.1-terra
Benchmarks
| Benchmark | Score | Notes |
|---|---|---|
| Terminal-Bench 4.0 | 50.3% | OpenAI-reported at max reasoning effort; GPT-5.6 Terra scored 21.5% |
| SEC-Bench Pro | 83.33% | pass@1, OpenAI system card addendum |
| HealthBench Professional | 62.7 | Length-adjusted, OpenAI system card addendum; GPT-5.6 Terra scored 57.7 |
Source: OpenAI GPT-6.1 Terra launch post and system card addendum
Tools powered by GPT-6.1 Terra
Compare GPT-6.1 Terra
Compare GPT-6.1 Terra with any other model
Build a comparison →All model comparisons →