OpenAI launched GPT-6.1 Terra on October 26, 2026, across the API, ChatGPT Work, and Codex, claiming faster computer-use task completion. Its standard API token rates are 50% lower than those of GPT-5.6 Terra and GPT-6.1 Sol; this reduction applies to API usage rather than ChatGPT subscription prices or total operating costs. OpenAI reports task completion more than three times faster than GPT-6 Astra and GPT-6.1 Sol; these are vendor-reported results, not independently established. This article examines the performance evidence, pricing qualifications, rollout availability, and considerations developers should weigh before choosing Terra.
GPT-6.1 Terra is an OpenAI model focused on visual reasoning, precise computer use, and writing. OpenAI describes work spanning proposals, technical drawings, forms, video-editing workflows, and invoice processing. The model combines interpreting visual information with performing computer-based tasks and producing written material.
The API model identifier is gpt-6.1-terra. It supports text and image input with text output, has a 1,050,000-token context window, and a maximum output of 128,000 tokens. Native audio and video inputs are not supported.
This distinction matters when interpreting references to video editing: Terra can operate within video-editing applications but does not accept native video files as input. Its visual-reasoning capabilities relate to supported image inputs rather than every visual medium. The practical difference lies between the workflows Terra assists with and the model's supported input and output formats.
The 2026 launch is significant because Terra combines lower standard API token prices with faster computer-use task completion, according to OpenAI. Terra completes tasks more than three times faster than GPT-6 Astra and GPT-6.1 Sol, with standard API token prices 50% lower than those of GPT-5.6 Terra and GPT-6.1 Sol. This pricing applies to API usage, not ChatGPT subscription costs.
OpenAI positions Terra for workflows involving visual reasoning, precise computer use, and writing, including preparing proposals, interpreting technical drawings, completing forms, processing invoices, and video-editing workflows. These examples focus on office work, extending beyond text generation alone.
For developers targeting such workflows, Terra's reported speed and pricing could merit evaluation where responsiveness and token costs are important. However, these are vendor-reported claims without independent customer outcome data demonstrating direct productivity gains.
OpenAI positions GPT-6.1 Terra around visual reasoning and precise computer use, emphasizing several evaluation areas based on vendor-reported improvements rather than proven customer pain points.
Problem 1: Task-completion speed. Terra reportedly completes tasks more than three times faster than GPT-6 Astra and GPT-6.1 Sol. Teams should verify if this speedup applies to their specific workflows.
Problem 2: Visual-workflow handling. Terra targets detailed visual reasoning and precise computer use for tasks like reading technical drawings, filling out forms, editing videos, and processing invoices. However, no separate visual benchmark results are provided.
Problem 3: Factual errors. OpenAI’s internal difficult-prompt evaluation shows factual error rates dropping from 53.6% with GPT-5.6 Terra to 16.7% with Terra at low reasoning effort. This is an internal test, not independently validated for real-world accuracy.
Problem 4: Complex coding. Terra scored 50.3% on Terminal-Bench 4.0 at maximum reasoning effort, compared with GPT-5.6 Terra’s 21.5%. Despite this, OpenAI recommends GPT-6.1 Sol or GPT-6 Astra for the most complex coding tasks.
These challenges align with Terra’s focus but require teams to distinguish OpenAI’s reported gains from demonstrated performance in their own environments.
Developers should review billing conditions, API requirements, and access restrictions alongside Terra’s standard token prices. OpenAI specifies five key selection criteria.
gpt-6.1-terra).Terra’s fit depends on workload and deployment setup; token pricing alone does not determine cost or availability in specific environments.
OpenAI positions Terra for practical work combining writing, visual reasoning, and precise computer use. The announcement identifies these application areas, without named customer deployments or detailed outcome data.
These use cases reflect intended applications rather than proven results. OpenAI continues to recommend GPT-6.1 Sol or GPT-6 Astra for the most complex coding tasks, distinguishing Terra’s focus on everyday work from models targeted at harder programming challenges.
OpenAI recommends comparing GPT-6.1 Terra against its predecessor on representative images and workflows at equal reasoning effort. That comparison should sit alongside other practical checks drawn from OpenAI's model documentation.
These steps help developers make informed migration decisions based on documented capabilities and pricing.
OpenAI reports these vendor-supplied benefits for GPT-6.1 Terra; none guarantee equivalent results on every workflow.
Teams should interpret these findings as vendor-reported rather than independently verified evidence.
Terra's pitch is volume: if an agent can read a screen, fill in a form, or process an invoice at half the previous token price, companies can hand more routine desktop work to AI. The same shift is changing how people find products and services. As assistants like ChatGPT increasingly research, compare, and act on a user's behalf, the brands those agents surface become the brands that get considered.
That has pushed a growing number of companies toward generative engine optimization, the practice of tracking and improving how AI systems describe and cite a brand. Firms such as XLR8 AI work in this space, monitoring brand visibility across AI-generated answers. Faster, cheaper agents like Terra are likely to make that visibility matter more, not less, because more buying and research decisions will pass through an AI intermediary before a person ever visits a website.
Disclosure: Anupma Kumari is Growth Lead at XLR8 AI. XLR8 AI has no partnership with OpenAI, and this article draws only on publicly reported information about GPT-6.1 Terra.
GPT-6.1 Terra offers standard API token prices 50% below GPT-5.6 Terra and GPT-6.1 Sol combined with OpenAI-reported improvements in computer-use performance, including task completion more than three times faster than GPT-6 Astra and GPT-6.1 Sol. These pricing reductions apply to API tokens, not subscriptions; performance claims are vendor-reported and not guaranteed across all workloads.
Before migrating, readers should review official model documentation to confirm access, applicable pricing, and workspace permissions. They should also compare Terra with predecessor models on representative images and workflows at equal reasoning effort to assess fit. Terra is available in ChatGPT Work and Codex but not ordinary Chat, with workspace settings potentially affecting availability.
GPT-6.1 Terra is an OpenAI model designed for visual reasoning, precise computer use, and writing. It launched on October 26, 2026, across the API, ChatGPT Work, and Codex. Intended workflows include proposals, technical drawing interpretation, form completion, invoice processing, and video editing. The API identifier is gpt-6.1-terra. Terra is not available in ordinary Chat but through ChatGPT Work and Codex.
Developers may consider Terra because it integrates visual reasoning, precise computer use, and writing for practical workflows such as forms and invoice processing. OpenAI reports task completion more than three times faster than GPT-6 Astra and on-par benchmark performance at about one-tenth Astra’s per-token price. These vendor-reported metrics suggest Terra merits evaluation for specific automation needs.
OpenAI recommends GPT-6.1 Sol or GPT-6 Astra for the most complex coding tasks. Despite Terra’s improved Terminal-Bench 4.0 score (50.3% at maximum reasoning effort), Sol and Astra remain preferred for hardest programming workloads. Model choice depends on workload specifics.
Terra’s standard API rates are $1 per million input tokens, $0.05 per million cached input tokens, and $5 per million output tokens. These rates are 50% lower than GPT-5.6 Terra and GPT-6.1 Sol. Requests over 272,000 input tokens incur premiums on the full request; Fast mode and regional processing also have premium fees.
No. The 50% price reduction applies to Terra’s standard API token rates, not ChatGPT subscription prices. API billing and ChatGPT subscriptions are separate. Access to Terra through ChatGPT plans does not increase plan usage limits.
Terra is available on Plus, Pro, Business, Enterprise, and Edu plans, subject to region, plan experience, and workspace permissions. It is accessible via ChatGPT Work and Codex, not ordinary Chat. Workspace settings can limit availability even for eligible plans.
Terra’s benchmark results are OpenAI-reported and have not been independently replicated. Internal evaluations show factual error rates decreasing significantly, but these should not be assumed as real-world error rates.
No. Terra’s API supports text and image input with text output; native audio and video inputs are unsupported. Video-editing workflow references relate to operating software, not accepting native video files.
