Loading rankings…
OpenAI
Base family: GPT-5.3 Codex
General intelligence · Rank 59 · Score (source scale) 33 · Leaderboard / snapshot date 2026-10-02
Estimated; not independently benchmarked
Official site ↗GPT-5.3-Codex 是 OpenAI 为 Codex 等环境中的智能体编程任务优化的模型,支持文本与图像输入及文本输出。
OpenAI 将其描述为面向智能体编程的模型,并允许 low、medium、high、xhigh 推理 effort;官方型号页未提供此处可引用的任务通过率或与同测竞品的质量对照,故仅据其明确任务定位描述,不作更广泛的代码质量保证。
Multimodal, suited to combined-content tasks such as Inputs: Text and Image; Outputs: Text, Code, and Tool calls.
Profile checked: 2026-10-03