OpenAI has released GPT-6.1 Terra, a new model focused on computer-use and vision capabilities. According to a report by MarkTechPost, the release comes with a significant price reduction for developer API usage, lowering the cost of agent tasks compared to previous models like GPT-6.1 Sol and GPT-6 Astra.
API Pricing and Performance Metrics
The new model is priced at one dollar per million input tokens and five dollars per million output tokens. In benchmark testing cited in the coverage, GPT-6.1 Terra achieved a medium-effort score on Agents’ Last Exam matching GPT-6 Astra, while completing the tasks at a substantially lower cost per task.
Open Questions and Missing Details
While the initial cost and benchmark figures are available, the announcement did not specify a full rollout timeline or availability details for all enterprise tiers. Furthermore, technical specifications regarding latency variations, exact context window limits, and fine-tuning support remain unaddressed at this time.
Developers and enterprise users will need to monitor OpenAI’s documentation updates to determine how GPT-6.1 Terra integrates into existing agentic workflows.
Lead image: panumas nikhomkhai / Pexels
