GPT-6.1 Terra
GPT-6.1 Terra is OpenAI's fast model for visual reasoning, computer use, and writing, released October 26, 2026.
Provider
OpenAI
Model family
OpenAI GPT
Multimodal LLM
Cost tier
Terra
Status
Current
Release Oct 26, 2026
Why teams choose it
When moving from GPT-5.6 Terra, OpenAI says to compare both models on the same images
and the same reasoning effort.
Codex CLI: codex -m gpt-6.1-terra
The Codex models page recommends Terra for everyday writing
images, and computer use, and keeps Astra for the hardest end-to-end work.
Tradeoffs to know
- Tool calling requires the Responses API.
- Live, Realtime, Assistants, fine-tuning, embeddings, image generation, video, and speech endpoints are not supported.
- Prompts over 272,000 input tokens are billed at the long-context rates for the whole request.
Benchmarks
No independent or vendor benchmark scores in this catalog entry.
Technical specs
- Inputs
- text, image
- Outputs
- text
- License
- Proprietary API
- API ID
gpt-6.1-terra- Context tokens
- 1,050,000
- Default Effort
- medium
- Knowledge Cutoff
- 2026-06-01
- Max Input Tokens
- 922,000
- Max output tokens
- 128,000
- Input Usd Per Million
- 1
- Output Usd Per Million
- 5
- Cache Read Usd Per Million
- 0.05
- Long Context Input Multiplier
- 2
- Long Context Output Multiplier
- 1.5
- Cache Write 12h Usd Per Million
- 2.5
- Cache Write 30m Usd Per Million
- 1.25
- Long Context Threshold Input Tokens
- 272,000
- Capabilities
- Text and image input with text output
- 1,050,000-token context window and 128,000 max output
- Computer use, vision, and writing on the Responses API
- Available in the API, ChatGPT Work, and Codex CLI 0.159.0 or later
Sourcehttps://developers.openai.com/api/docs/models/gpt-6.1-terra
OpenAI GPT family lineup
Current models
Previous versions
Related models
GPT-6.1 Terra FAQ
What is GPT-6.1 Terra?
1 Terra is OpenAI's fast model for visual reasoning, computer use, and writing, released October 26, 2026. 1-terra. It takes text and images and returns text, with a 1,050,000-token context window, 922,000 max input tokens, 128,000 max output tokens, and a knowledge cutoff of June 1, 2026. effort accepts none, low, medium (default), high, xhigh, and max. Tool calling requires the Responses API. 05 cached input, and $5 output per million tokens. 50 for 12-hour retention, before long-context adjustments. 5x output for the full request. Batch and Flex are 50% of Standard. Regional processing adds 10%. Fast mode is 2x Standard. 0, ChatGPT Credits, and the API.
When does GPT-6.1 Terra fit best?
Match product photos to catalog entries
What should teams watch out for with GPT-6.1 Terra?
Tool calling requires the Responses API.
Explore next
Models, tools, and comparisons that connect to this reference.