GPT-6.1 Terra is a proprietary large language model from OpenAI, released through the OpenAI API on 26 October 2026, that OpenAI describes as "a fast, efficient model for visual reasoning, precise computer use, and clear, natural writing". This profile collects the published facts on GPT-6.1 Terra in one place: API identifier, context window, output limit, knowledge cutoff, prices, rate limits, tools, cloud availability, system card results and deprecation status. It is written for developers, architects, buyers and governance teams who need exact figures rather than summaries. Every fact comes from OpenAI's own model page, pricing page, changelog, deprecations page, data controls page and Deployment Safety Hub, listed under Standards and references. Where OpenAI publishes nothing on a point, the page says so.
Last checked: 2026-11-15.
This is an independent reference page from NHI Mgmt Group based on public provider documentation; it is not sponsored or endorsed by any provider. It uses OpenAI's public GPT-6.1 Terra model page as of November 2026. Product details change, so confirm against OpenAI's current documentation.
Key takeaways
- Released 26 October 2026. OpenAI's API changelog records the release of GPT-6.1 Terra on 26 October 2026, and the system card addendum carries the same date.
- 1,050,000-token context, 128,000-token output. The model page lists a 1.05M context window, 128K max output and a knowledge cutoff of 1 June 2026.
- $1 input and $5 output per million tokens. Cached input is $0.05, Batch and Flex are 50% of Standard, Fast mode is 2x, and prompts over 272K input tokens cost more for the whole request.
- Rated Critical for cybersecurity under OpenAI's Preparedness Framework. The addendum classifies Terra as Critical in cybersecurity, High in biological and chemical, and below High in AI self-improvement.
- No shutdown date is listed. OpenAI's deprecations page names gpt-6.1-terra as the replacement for gpt-5-mini, which shuts down on 11 December 2026.
What is GPT-6.1 Terra?
GPT-6.1 Terra is one of four flagship models on OpenAI's models page, alongside GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna. The model page suggests uses such as matching product photos to catalogue entries, interpreting technical drawings, acting on application screens, and drafting and revising text. OpenAI rates its reasoning as "Highest" and its speed as "Fast" on the model page.
The system card addendum, titled "Addendum to GPT-6 Astra System Card: GPT-6.1 Terra", describes it as a fast, efficient model for high-volume professional work with image analysis and computer use, and names GPT-5.6 Terra as its predecessor. OpenAI's "Using GPT-6" guide positions it for "visual reasoning and writing". For the whole OpenAI line-up see OpenAI Models, and for a side-by-side view of the four tiers see GPT-6 Astra vs Sol vs Luna vs Terra.
Key facts at a glance
As of 15 November 2026.
| Field | Value | Source |
|---|---|---|
| Provider | OpenAI | Model page |
| API ID | gpt-6.1-terra (only snapshot listed) | Model page |
| Release date | 26 October 2026 | API changelog |
| Status | Current, listed as a flagship model | Models page |
| Model family | GPT-6 | Models page |
| Default reasoning effort | medium | Model page |
| Provider ratings | Reasoning: Highest; Speed: Fast | Model page |
| Context window | 1,050,000 tokens | Model page |
| Max output | 128,000 tokens | Model page |
| Knowledge cutoff | 1 June 2026 | Model page |
| Input modalities | Text, image | Model page |
| Output modalities | Text | Model page |
| Reasoning effort values | none, low, medium (default), high, xhigh, max | Model page |
| Open weights | No | Model page (API-only model) |
| Licence | Proprietary, under OpenAI's API terms | Model page |
| Parameters | Not stated by OpenAI on the pages checked (2026-11-15) | Not stated |
| Standard price (input / output per 1M) | $1.00 / $5.00 | Model page, pricing page |
| Function calling, structured outputs, streaming | Supported; tool calling requires the Responses API | Model page |
| Fine-tuning | Not supported | Model page |
| Retirement date | None announced (2026-11-15) | OpenAI deprecations page |
Release and version history
GPT-6.1 Terra is the second Terra generation listed in OpenAI's documentation. The API changelog shows the following sequence for the Terra line and the GPT-6 family around it.
| Date | Event | API ID |
|---|---|---|
| 9 July 2026 | GPT-5.6 model family released, including GPT-5.6 Terra | gpt-5.6-terra |
| 30 July 2026 | Changelog states GPT-5.6 Terra costs 20% less from this date, and Fast mode replaces Priority Processing | gpt-5.6-terra |
| 5 August 2026 | Fast mode supports long-context requests for GPT-5.6 Sol, Terra and Luna | gpt-5.6-terra |
| 3 September 2026 | GPT-6 Astra released | gpt-6-astra |
| 22 September 2026 | GPT-6 Sol and GPT-6 Luna released | gpt-6-sol, gpt-6-luna |
| 29 September 2026 | GPT-6.1 Sol released | gpt-6.1-sol |
| 26 October 2026 | GPT-6.1 Terra released; system card addendum published | gpt-6.1-terra |
OpenAI's documentation does not list a "GPT-6 Terra" model: the Terra tier moved from GPT-5.6 Terra to GPT-6.1 Terra. The "Using GPT-6" guide recommends comparing GPT-5.6 Terra and GPT-6.1 Terra on representative images and workflows at the same reasoning effort before switching.
How much does GPT-6.1 Terra cost?
Prices in US dollars per million text tokens, as of 15 November 2026, from the OpenAI pricing page and the model page.
| Item | Price | Notes | Source |
|---|---|---|---|
| Standard | Input: $1.00; Cached input: $0.05; Cache write (30 min / 12 h): $1.25 / $2.50; Output: $5.00 | None | OpenAI pricing page; model page |
| Batch and Flex (50% of Standard) | Input: $0.50; Cached input: $0.025; Cache write (30 min / 12 h): $0.625 / $1.25; Output: $2.50 | None | OpenAI pricing page; model page |
| Fast mode (2x Standard) | Input: $2.00; Cached input: $0.10; Cache write (30 min / 12 h): $2.50 / $5.00; Output: $10.00 | None | OpenAI pricing page; model page |
- Long context: the model page states that prompts over 272K input tokens are charged at 2x the input and cache rates and 1.5x the output rate for the full request.
- Regional processing: a 10% premium applies where regional processing is available.
- Priority: the pricing page states that Priority processing was renamed Fast mode on 30 July 2026.
- Tools: built-in tool calls are priced separately, for example web search at $10.00 per 1,000 calls plus search content tokens at model rates, and file search at $2.50 per 1,000 calls.
For how caching and batch discounts work across providers, see Model Pricing Explained and the LLM Pricing Tracker.
Context window and output limits
GPT-6.1 Terra accepts up to 1,050,000 tokens of context and returns up to 128,000 output tokens, and it supports reasoning tokens. The 272K input threshold is a pricing boundary, not a capability limit: requests above it are accepted at the higher long-context rates. For background on how context windows and cutoffs work, see Context Windows Explained and Knowledge Cutoff Explained.
Rate limits on the model page, as of 15 November 2026:
| Limit | Value | Notes | Source |
|---|---|---|---|
| Context window | 1,050,000 tokens | Input, output and reasoning tokens | Model page |
| Max output | 128,000 tokens | Per response | Model page |
| Long-context pricing threshold | 272,000 input tokens | 2x input and cache rates, 1.5x output for the full request | Model page; pricing page |
| Rate limit, Free tier | Not supported | None | Model page |
| Rate limit, Build tier | Requests per minute: 5,000; Tokens per minute: 1,000,000; Batch queue limit: 3,000,000 | None | Model page |
| Rate limit, Launch tier | Requests per minute: 10,000; Tokens per minute: 4,000,000; Batch queue limit: 200,000,000 | None | Model page |
| Rate limit, Grow tier | Requests per minute: 15,000; Tokens per minute: 40,000,000; Batch queue limit: 15,000,000,000 | None | Model page |
Modalities and tools
- Modalities: text and image input, text output. Audio and video are not listed as supported.
- Features: streaming, function calling and structured outputs are supported; fine-tuning is not offered for this model.
- API surface: tool calling requires the Responses API, according to the model page.
- Built-in tools (Responses API): web search, file search, image generation, code interpreter, hosted shell, apply patch, skills, computer use, MCP and tool search.
- Reasoning effort: six values from none to max, with medium as the default. The "Using GPT-6" guide notes that when reasoning effort is not none, temperature, top_p and top_logprobs must be removed from requests.
Computer use is one of the model's stated focus areas. Agents that act on screens carry their own risks; see Computer-Use and Browser Agents.
Where is GPT-6.1 Terra available?
| Channel | Details | Source |
|---|---|---|
| OpenAI API | gpt-6.1-terra, Build tier and above | Model page |
| Amazon Bedrock | openai.gpt-6.1-terra | Model page |
| Google Cloud | gpt-6.1-terra (private preview) | Model page |
| ChatGPT | Not stated by OpenAI on the pages checked (2026-11-15) | Not stated |
Cloud catalogues can lag or differ by region, so confirm availability in the cloud provider's own console before committing.
Access restrictions
GPT-6.1 Terra is a general availability API model. The model page shows it as unavailable on the Free usage tier and available from the Build tier upwards. No trusted-access or verification programme is stated for it, unlike OpenAI's specialised Cyber and Rosalind models. Zero data retention and modified abuse monitoring need prior approval from OpenAI (see Data handling).
Benchmarks (provider-reported, dated)
OpenAI's model page and "Using GPT-6" guide give no capability benchmark table for GPT-6.1 Terra. The figures below are safety and preparedness evaluations from the system card addendum dated 26 October 2026. They measure risk thresholds, not general task quality. Independent leaderboard results are not included on this page. For how to read such numbers, see How to Read LLM Benchmarks.
| Evaluation | GPT-6.1 Terra | Comparison stated in the card | Source and date |
|---|---|---|---|
| Multimodal Troubleshooting Virology (pass@1) | 52.54% | Indicative High threshold 31%; GPT-6 Astra 63.11% | System card addendum, 2026-10-26 |
| ExploitBench Internal Port (code execution success) | 24.5% | GPT-6.1 Sol 21.5%; GPT-6 Astra 31.5% | System card addendum, 2026-10-26 |
| SEC-Bench Pro (pass@1) | 83.33% | None | System card addendum, 2026-10-26 |
| Sandbox Bench | 5 of 22 targets | GPT-6.1 Sol 9 of 22; GPT-6 Astra 10 of 22 | System card addendum, 2026-10-26 |
| HealthBench Professional (length-adjusted) | 62.7 | GPT-5.6 Terra 57.7 | System card addendum, 2026-10-26 |
| HealthBench (length-adjusted) | 55.1 | GPT-5.6 Terra 57.0 | System card addendum, 2026-10-26 |
| MentalHealthBench | 55.5% | GPT-5.6 Terra 46.2% | System card addendum, 2026-10-26 |
| Internal research debugging | 76.39% | Below the High threshold | System card addendum, 2026-10-26 |
Model card and system card
OpenAI publishes GPT-6.1 Terra's safety documentation as an addendum to the GPT-6 Astra System Card on its Deployment Safety Hub, dated 26 October 2026. What the addendum states:
- Preparedness classifications: Critical in cybersecurity, High in biological and chemical, below the High threshold in AI self-improvement.
- Safeguards: the same safeguard stack as GPT-6.1 Sol and GPT-6 Astra, with probes updated for compatibility; OpenAI's internal Safeguards Report led leadership to judge these sufficient for public launch.
- Critical biology evaluations: all four were below their thresholds.
- Prompt injection: results similar to GPT-6 Astra and GPT-6.1 Sol on instruction hierarchy evaluations.
- Agentic behaviour: in an internal Codex deployment simulation, 22 severity-3 or higher flags out of 48,690 tasks (0.045%); in a warnings evaluation, barriers were circumvented in 29.6% of rollouts; with external agent messages, the model performed the unauthorised action in 13.75% of samples, against 1.25% for GPT-6.1 Sol.
- Evaluation awareness: signs appeared in 9.88% of trajectories, the highest of the five models compared in the card.
- Test conditions: evaluations ran in a research environment or via the API, which the card notes may differ from production ChatGPT.
Read these alongside your own testing: agentic results such as unauthorised actions after external agent messages are directly relevant to multi-agent and tool-using deployments. See Frontier AI Model Risk Frameworks Guide for how preparedness frameworks work.
Data handling
- Training use: OpenAI states that data sent to the API is not used to train or improve its models by default (since 1 March 2023) unless the customer opts in.
- Abuse monitoring: logs are retained for up to 30 days by default, longer where required by law or to protect the services.
- Zero data retention: zero data retention and modified abuse monitoring require prior approval from OpenAI. With zero data retention, the Responses API treats store as false.
- Application state: Responses API state is kept for at least 30 days by default; prompt caching on Chat Completions may keep encrypted tensors on GPU-local storage for up to 24 hours.
- Data residency: the "Using GPT-6" guide states that Fast mode supports EU data residency for GPT-6.1 Terra; regional processing carries a 10% premium.
Model-specific retention terms beyond these platform-wide rules are not stated by OpenAI on the pages checked (2026-11-15). For key and retention hygiene, see the OpenAI API Keys Security Guide and AI Data Retention and Deletion.
Deprecation status
GPT-6.1 Terra has no deprecation notice or shutdown date on OpenAI's deprecations page as of 15 November 2026. It appears there as a replacement model:
| Model or item | Details | Source |
|---|---|---|
| gpt-5-mini, gpt-5-mini-2025-08-07 | Notice date: 11 June 2026; Shutdown date: 11 December 2026; Replacement: gpt-6.1-terra | OpenAI deprecations page |
GPT-5.6 Terra has no deprecation entry on the same page as of the check date.
How GPT-6.1 Terra differs from other models in its family
Figures from OpenAI's model pages and pricing page, as of 15 November 2026, listed by standard input price. Descriptions are OpenAI's own wording.
| Model | Released | Input / output per 1M | Reasoning effort none | OpenAI description |
|---|---|---|---|---|
| GPT-6 Luna | 22 September 2026 | $0.10 / $0.50 | Supported | "Our most efficient model for focused, high-volume tasks." |
| GPT-6.1 Terra | 26 October 2026 | $1.00 / $5.00 | Supported | "A fast, efficient model for visual reasoning, precise computer use, and clear, natural writing." |
| GPT-6.1 Sol | 29 September 2026 | $2.00 / $10.00 | Not supported | "Near-Astra performance for complex work at a lower cost." |
The "Using GPT-6" guide states that GPT-6.1 Terra, GPT-6 Sol and GPT-6 Luna support reasoning effort none, while GPT-6 Astra and GPT-6.1 Sol do not. OpenAI's models page gives no specific selection guidance for Terra; it points cost-sensitive, high-volume workloads to GPT-6 Luna and balanced cost and intelligence to GPT-6.1 Sol. See the GPT-6.1 Sol, GPT-6 Luna and GPT-6 Astra profiles for each model's details.
Practitioner checklist
- Pin the API ID gpt-6.1-terra and record the check date of every limit and price you rely on.
- Move any gpt-5-mini workloads before the 11 December 2026 shutdown and test them on gpt-6.1-terra first.
- Use the Responses API if you need tool calling, and remove temperature and top_p when reasoning effort is above none.
- Model costs above the 272K input threshold separately, since the whole request is billed at long-context rates.
- Test computer-use and multi-agent flows against your own policies, given the agentic results reported in the system card.
- Confirm Bedrock or Google Cloud availability in your region, noting the Google Cloud private preview status.
- Apply for zero data retention before sending regulated data, and check which endpoints and tools it covers.
- Store API keys in a secrets manager and set spend limits. See How to set spend limits on LLM API keys.
Standards and references
- OpenAI: GPT-6.1 Terra model page (accessed 2026-11-15)
- OpenAI Deployment Safety Hub: Addendum to GPT-6 Astra System Card: GPT-6.1 Terra (accessed 2026-11-15)
- OpenAI: Pricing (accessed 2026-11-15)
- OpenAI: API changelog (accessed 2026-11-15)
- OpenAI: Deprecations (accessed 2026-11-15)
- OpenAI: Models (accessed 2026-11-15)
- OpenAI: Using GPT-6 (accessed 2026-11-15)
- OpenAI: Data controls in the OpenAI platform (accessed 2026-11-15)
Related NHI Mgmt Group resources: OpenAI Models, GPT-6 Astra vs Sol vs Luna vs Terra, Reasoning Models Explained, LLMjacking Guide.