OpenAI Releases GPT-6.2 Astra Model, After Previously Withdrawing the Previous Update
English text is being prepared.
OpenAI has released GPT-6.2 Astra, a new version of its most advanced model Astra, nearly two months after the previous update was withdrawn. OpenAI says the model is more efficient at performing highly difficult professional tasks, and that the slides, documents, and spreadsheets it generates better align with user instructions and industry standards, reading less like AI-generated content. In a benchmark test of multi-step professional work, it achieved better performance at a cost far lower than Anthropic's most powerful Fable model.
Why read this
SelectedAI score 76
The report explains the efficiency and cost comparison of the new Astra version on professional tasks, which can be used to understand the direction of OpenAI's model iteration in complex office scenarios.
Topics
Body
OpenAI正在发布其最先进的Astra人工智能模型新版本GPT-6.2 Astra。此前,该公司为建立更强的技术安全保障,撤回了上一版更新,距今已近两个月。 该公司周二表示,GPT-6.2 Astra旨在成为其应对复杂专业项目能力最强的模型,在执行高难度任务时效率更高,其生成内容“读起来更不像AI写的”。 OpenAI称,该模型生成的幻灯片、文档和电子表格更贴合用户指令和行业标准,且“读起来更不像AI生成的”。该公司还表示,最新版Astra在一项多步骤专业工作基准测试中,以远低于对手Anthropic PBC最强大的Fable模型的成本,取得了更优表现。风险提示及免责条款
市场有风险,投资需谨慎。本文不构成个人投资建议,也未考虑到个别用户特殊的投资目标、财务状况或需要。用户应考虑本文中的任何意见、观点或结论是否符合其特定状况。据此投资,责任自负。
AI Trends