OpenAI’s New Model Solves Complex Math Problems
OpenAI 的新模型解決了複雜的數學問題
更新於: 2026年8月2日 上午02:30
OpenAI is transforming the field of artificial intelligence with its latest breakthroughs in mathematical problem-solving.
OpenAI 正以其在數學問題解決領域的最新突破，徹底改變人工智慧領域。
While the 2024 'o1' series introduced 'chain-of-thought' reasoning—allowing the AI to think through problems step-by-step—the upcoming 'Astra' model family marks a significant leap forward.
儘管 2024 年的「o1」系列引入了「思維鏈」（Chain-of-Thought）推理——讓 AI 能夠逐步思考問題——即將推出的「Astra」模型系列則標誌著重大的飛躍。
Announced in August 2026, Astra is designed for deep, autonomous research.
Astra 於 2026 年 8 月發布，旨在進行深度的自主研究。
Unlike its predecessors, which focused on standardized tests, Astra has successfully tackled ten major mathematical and computer science problems that had remained unsolved for over a decade.
與專注於標準化測驗的前代產品不同，Astra 已成功攻克了十個十多年來懸而未決的重大數學和電腦科學問題。
To ensure absolute accuracy, these solutions were formalized in 'Lean,' a machine-verified language, and documented in a 249-page manuscript.
為了確保絕對的準確性，這些解決方案採用了 Lean進行形式化，並記錄在一份 249 頁的草稿中。
Although running these complex calculations is resource-intensive—costing roughly $2,000 per solution—Astra demonstrates that AI is becoming a powerful partner for human researchers.
儘管運行這些複雜計算需要消耗大量資源——每個解決方案成本約為 2,000 美元——但 Astra 證明了 AI 正成為人類研究人員強大的合作夥伴。
By moving toward autonomous, long-term research capabilities, OpenAI is pushing the boundaries of what computers can contribute to the frontier of human knowledge, marking a new era where AI actively advances scientific understanding.
透過邁向自主的長期研究能力，OpenAI 正在突破電腦在人類知識前沿所能貢獻的界限，標誌著一個 AI 主動推動科學理解的新時代。
