GPT-5.3上线Codex！OpenAI回应Claude新模型只用了15分钟

Core Viewpoint - The article discusses the release of OpenAI's latest programming model, GPT-5.3-Codex, which competes with Anthropic's Claude Opus 4.6, highlighting significant improvements in coding capabilities and user interface design [1][13]. Group 1: Model Improvements - GPT-5.3-Codex shows enhanced aesthetic appeal and design in its demos, including a racing game and a diving game [2][6]. - The model has demonstrated superior performance in various benchmarks, achieving 57% in SWE-Bench Pro, 76% in TerminalBench 2.0, and 64% in OSWorld [11][12]. - It has improved efficiency, requiring less than half the tokens compared to its predecessor for the same tasks, with a speed increase of over 25% [11][22]. Group 2: Functional Capabilities - GPT-5.3-Codex excels in computer use, assisting financial professionals in creating presentations and handling complex tasks like document writing and spreadsheet management [9][23]. - The model's ability to self-accelerate during its training process marks a significant advancement, allowing it to monitor and debug its own training tasks [28][29]. Group 3: Business Applications - OpenAI is launching Frontier, a platform aimed at integrating AI into corporate workflows, with notable companies like HP and Uber already adopting it [34][38]. - The AI4S initiative, in collaboration with Ginkgo, aims to reduce protein synthesis costs by 40% using GPT-5, showcasing the model's application in synthetic biology [39][41].