Gemini 3 Deep Think 升级：推动科学、研究与工程前沿
原标题：Gemini 3 Deep Think: Advancing science, research and engineering
AI 摘要
Google DeepMind 发布了 Gemini 3 Deep Think 推理模式的重要升级，该模式专为科学、研究和工程领域设计，可处理缺乏明确规则和数据的复杂问题。新版 Deep Think 在多个基准测试中创下新高，如 Humanity's Last Exam 得分 48.4%、ARC-AGI-2 得分 84.6%、Codeforces Elo 3455，并在国际数学和物理、化学奥林匹克中达到金牌水平。该模式现已面向 Google AI Ultra 订阅者开放，并首次通过 Gemini API 向选定的研究人员、工程师和企业提供早期访问。
正文节选
Gemini 3 Deep Think: Advancing science, research and engineering Today, we’re releasing a major upgrade to Gemini 3 Deep Think, our specialized reasoning mode, built to push the frontier of intelligence and solve modern challenges across science, research, and engineering. We updated Gemini 3 Deep Think in close partnership with scientists and researchers to tackle tough research challenges — where problems often lack clear guardrails or a single correct solution and data is often messy or incom