OpenAI: GPT-5.1-Codex-Mini

GPT-5.1-Codex-Mini is a text model from OpenAI, released on 13 November 2025. It is a smaller and faster version of GPT-5.1-Codex, positioned within the GPT-5.1 family of coding-focused models. The provider description frames it as the lighter sibling of the full Codex model, which is the main thing to keep in mind when deciding between the two.

The model accepts image and text input and returns text output. It has a context window of 400,000 tokens and a maximum output of 128,000 tokens, so it can hold very large inputs (long source files, multiple files, or extended transcripts) while still producing substantial responses. Supported request parameters include reasoning controls (reasoning, reasoning_effort, include_reasoning), max_completion_tokens, response_format, structured_outputs, seed, tool_choice, tools, and verbosity.

Those capabilities point to code-centric work: reading and reasoning over large repositories or long files, generating or refactoring code, answering questions about an existing codebase, and handling tasks where an image is part of the input, such as a screenshot of an interface, a diagram, or an error. Because tools and tool_choice are supported, it can also drive function calling and agent-style workflows, and structured_outputs plus response_format make it suitable when you need responses in a fixed schema rather than free text. The reasoning parameters let you trade latency for more deliberate problem solving on harder tasks.

You call it through api14's OpenAI-compatible API using the model id "openai/gpt-5.1-codex-mini". Because the endpoint follows the OpenAI chat completions format, existing OpenAI SDKs and clients work with only a base URL and key change. In the request you can set any of the supported parameters listed above, including tools and tool_choice for function calling, response_format or structured_outputs for schema-constrained output, seed for more reproducible sampling, and reasoning_effort or verbosity to steer how much internal reasoning and how much output detail you get.

The main trade-off is that this is the mini variant: it is built to be smaller and faster than GPT-5.1-Codex, so for the most demanding coding or reasoning tasks the full model may be the better fit, while the mini version suits high-volume or latency-sensitive work. It is also text-output only, so if you need generated images or audio, this model is not the right choice.