OpenAI Unveils GPT-6.1 Terra, Boosting AI Efficiency

"We’re introducing GPT-6.1 Terra, a fast, efficient model that combines detailed visual reasoning and precise computer use alongside clearer, more natural writing," OpenAI explained in a blog post.

What Terra is built to do

OpenAI positioned GPT-6.1 Terra as a model tailored to everyday, computer-based work rather than the broadest or most compute-intensive applications. According to the company, Terra is meant to speed and lower the cost of routine tasks such as filling out forms, editing videos, interpreting technical drawings, and automating repetitive desktop work. OpenAI framed the release as "a significant step up from GPT-5.6 Terra" with particular strengths in visual reasoning and precise control of software environments.

Speed, pricing, and product positioning

OpenAI set explicit price points and comparative claims for Terra. API access to GPT-6.1 Terra is priced at $1 per million input tokens and $5 per million output tokens — a 50% reduction versus GPT-5.6 Terra and GPT-6.1 Sol, the company said. On task speed, OpenAI reported that Terra completes tasks "over three times faster" than both GPT-6 Astra and GPT-6.1 Sol. The company also stated that Terra "performs on par with GPT-6 Astra at only one-tenth its price per token" in some comparisons, underscoring a cost-performance argument in favor of Terra for many routine workloads.

Benchmarks: agents, hallucinations, and coding

OpenAI provided benchmark results to support its claims. On "Agents' Last Exam," Terra scored 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort, while costing about one-third as much in that measurement. OpenAI also reported marked reductions in hallucination-type errors on its internal benchmark: the error rate at low reasoning effort fell from 53.6% with GPT-5.6 Terra to 16.7% with GPT-6.1 Terra. At maximum reasoning effort, OpenAI said the error rate decreased by 18 percentage points to 4.3%, a level the company described as "nearly matching the rates for GPT-6 Astra and GPT-6.1 Sol."

Coding performance improved too. On Terminal-Bench 4.0 at maximum reasoning effort, Terra scored 50.3%, more than double GPT-5.6 Terra's 21.5%. OpenAI nevertheless recommended GPT-6.1 Sol or GPT-6 Astra for the most complex coding tasks, signaling a deliberate product differentiation: Terra for fast, cost-efficient general work and visual reasoning, Sol/Astra for peak coding complexity.

Availability and named access paths

GPT-6.1 Terra is available immediately to specific user tiers and products. OpenAI listed Plus, Pro, Enterprise, and Edu users as having access via ChatGPT Work and Codex, and the model is exposed through the API as "gpt-6.1-terra." That distribution strategy places Terra inside both consumer-facing and enterprise-facing product lines while making the model callable programmatically for integrations and automation.

What this means for technologists, procurement leaders, and end users

• Technologists and security teams: Expect to validate Terra's claims against real workflows — for example, running Agents' Last Exam–type evaluations and Terminal-Bench 4.0 tests to confirm the reported gains in speed, error reduction, and coding quality. The company’s own benchmarks show large reductions in certain error rates, but teams will need to measure those improvements in their specific production contexts.
• Procurement and enterprise leaders: The published API pricing ($1 per million input tokens, $5 per million output tokens) and stated 50% cost advantage versus some earlier models provide a clear economic argument for shifting routine automation and visual-reasoning workloads to Terra. Availability to Plus, Pro, Enterprise, and Edu users also makes trial and staged rollouts straightforward.
• End users and workflow owners: If Terra’s claims hold up in practice, tasks such as form filling, video editing, and interpreting technical drawings could become materially faster and cheaper; OpenAI’s claim of completing tasks over three times faster than some larger models promises reductions in latency for interactive use and lower compute bills for repeated jobs.

OpenAI’s release frames GPT-6.1 Terra as a focused optimization: faster, cheaper, and tuned for day-to-day computer tasks, while leaving the highest-end coding and compute-heavy problems to Sol and Astra. The company’s stated benchmarks and specific pricing invite straightforward tests — and straightforward choices — about when to trade peak model capability for speed and cost. How widely organizations reassign those workloads to Terra will depend on how its reported improvements play out in real production environments.

Read the original OpenAI announcement: https://www.bleepingcomputer.com/news/artificial-intelligence/openai-launches-gpt-61-terra-with-3x-faster-ai-agents-50-percent-lower-pricing/