Introducing GPT-6.2 Astra and Sol
Summary
OpenAI introduces GPT-6.2 Astra, a frontier model for complex professional work, and GPT-6.2 Sol, an efficient model for everyday tasks. Both models deliver better results with fewer output tokens, lowering costs at scale. Astra sets new benchmarks on Agents' Last Exam (5.2 points above GPT-6 Astra at less than a third of the cost) and Terminal-Bench-Science 0.1, and excels at professional slides, spreadsheets, documents, legal work (Harvey LAB-AA), coding, cybersecurity, and 3D design. Sol comes within 0.5 points of Astra on FrontierCode 2.0 at about one-fifth the cost. API pricing: Astra input $8, cached input $0.40, output $40 per million tokens (20% lower than GPT-6 Astra with a 60% cache discount); Sol keeps GPT-6.1 Sol's rates ($2/$0.10/$10). Both meet the Critical cybersecurity threshold and High biological/chemical threshold under the Preparedness Framework; Astra is the first externally released model to meet the High threshold for AI Self-improvement. New safeguards, including Control Room and expanded monitoring, accompany the launch. Models are available to Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex, and via the API.
(Source:OpenAI)