gpt-5.3-codex

gpt-5.3-codex is OpenAI’s Codex model identifier for agent-based software engineering. The official model catalog uses the Codex family for long-chain coding tasks, such as understanding the code base, planning changes, editing multiple files, running tests, troubleshooting errors, and calling development tools. It is suitable for assisting R&D work under clear warehouse scope, test standards and change constraints. When calling, the project structure, language version, dependencies, tools allowed to be accessed, and unmodifiable areas should be provided, and the generated changes should be placed under version control, automated testing, and security review.

Context
400000 tokens
Max output
128000 tokens
Input
llm image
Output
llm

Pricing

Input: $1.75 / 1M tokens

Output: $14 / 1M tokens

Cached input: $0.17500000000000002 / 1M tokens

Endpoints

  • openai

More models

  • gpt-4o

    gpt-4o 是 OpenAI GPT-4o 系列模型标识。官方目录将 GPT-4o 描述为快速、灵活的通用模型,mini 版本则用于更聚焦、成本更低的任务。它们可用于文本处理、图像理解、问答、摘要、分类和应用内助手等工作流。选型时应根据响应质量、延迟、输入模态和吞吐量区分标准版与 mini 版,并针对中文、表格、图片和长文档分别测试。重要事实和高风险建议仍需由检索结果或人工审核确认。

  • grok-4.7
  • vertex/gemini-3.5-flash-lite
  • gpt-6-astra-1009