OpenAI Updates GPT-6 Astra Interface, Adds Information on Model Performance and API Pricing

According to reports from Cailian Press, OpenAI has updated the GPT-6 Astra product page, adding details about model performance, enterprise applications, security capabilities, and API pricing. Astra achieved 99.9% on ARC-AGI-3, approximately 97.6% on FrontierMath Tier 4, and 57.9% on Terminal-Bench 4.0, surpassing GPT-5.6 Sol's 37.3% and slightly exceeding Claude Fable 5.1's 55.8%. On the professional task Agents’ Last Exam, Astra scored 59.3%, also higher than Claude Opus 5's 55.5%.