OpenAI: GPT-5.4 Mini

GPT-5.4 Mini is a text model from OpenAI, positioned as a faster, more efficient member of the GPT-5.4 family. It brings the core capabilities of GPT-5.4 to workloads where throughput matters, which makes it a reasonable default when you want GPT-5.4 style behaviour without the latency or capacity profile of the larger model. It was released by the provider on 2026-03-17, and its knowledge cutoff is 2025-08-31.

The model accepts text, image and file inputs, and it returns text. Its context window is 400,000 tokens, with a maximum output of 128,000 tokens, so long documents, large codebases or extended conversation histories can be handled in a single request. The provider describes strong performance across reasoning and coding, and the model exposes reasoning controls, which means you can tune how much internal reasoning effort it applies rather than treating it as a fixed black box.

Those capabilities point to a fairly wide set of practical uses. Document and file analysis works well given the combination of file input and a large context: contracts, reports, logs or research material can be passed in and summarised or queried. Image input supports tasks such as reading screenshots, diagrams or scanned pages alongside text instructions. Coding assistance, structured data extraction and multi step reasoning over long inputs are all natural fits, and the reasoning effort and verbosity parameters let you trade depth against speed depending on the task.

You call it through api14's OpenAI-compatible API using the model id "openai/gpt-5.4-mini". Because the interface follows the OpenAI chat completions shape, existing OpenAI SDKs and clients generally work with only a base URL and key change. The supported request parameters are include_reasoning, max_completion_tokens, max_tokens, reasoning, reasoning_effort, response_format, seed, structured_outputs, tool_choice, tools and verbosity. In practice that means you can cap output length, request deterministic sampling with seed, enforce JSON style output through response_format or structured_outputs, define function calling with tools and tool_choice, and control reasoning and verbosity explicitly.

There is no stated limitation beyond the modality boundary: output is text only, so image or audio generation is out of scope. If your workload depends on non-text output, a different model is required. Otherwise, GPT-5.4 Mini is a sensible choice when you want a large context, image and file understanding, tool use and configurable reasoning in a single model.