OpenAI released GPT-6.1 Terra on October 26, 2026, a model that halves the standard API pricing of its predecessor for visual reasoning and computer-use tasks.
The price cut directly targets high-volume automation workloads where cost per task governs adoption. Standard access is priced at $1 per million input tokens and $5 per million output tokens, according to the company's announcement1. That is 50 percent below the rates for GPT-5.6 Terra and GPT-6.1 Sol.
Cached input tokens cost $0.05 per million, or 5 percent of the uncached rate, per the API documentation2. Batch and Flex processing carry a further 50 percent discount off standard prices. For prompts exceeding 272,000 input tokens, OpenAI doubles the input and cache rates and applies a 1.5x multiplier to output pricing for the full request. The announcement did not detail capacity available at launch.
The price cut brings the Terra line below competing models. One early social-media reaction noted Terra is now "half the price of Sonnet" @bindureddy on X4. OpenAI's own comparisons frame the model as roughly one-tenth the price per token of its frontier GPT-6 Astra on computer-use benchmarks.
OpenAI described the model as a fit for tasks such as reading invoices and entering data into accounting software. "Terra continues to advance the cost-intelligence frontier across our model families, providing a new balance of speed, capability, and cost," the company said.
On the Agents' Last Exam benchmark—which includes turning architectural drawings into 3D models and producing time-coded film logs—GPT-6.1 Terra scored 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort and about one-third the cost. At maximum effort, it completed tasks over three times faster than both GPT-6 Astra and GPT-6.1 Sol, the company said.
The model solved 50.3 percent of Terminal-Bench 4.0 tasks at maximum reasoning, more than double GPT-5.6 Terra's published 21.5 percent. On an internal factuality benchmark, the error rate at low reasoning effort fell from 53.6 percent to 16.7 percent.
OpenAI released the model alongside a system card detailing safety evaluations. The company said GPT-6.1 Terra meets the Critical threshold in cybersecurity under its Preparedness Framework. OpenAI did not provide independent third-party benchmarks or specify the compute infrastructure powering the model at scale.