OpenAI: GPT-5.4

GPT-5.4 is OpenAI's frontier text model, released on 5 March 2026. It unifies the Codex and GPT lines into a single system, so instead of choosing between a coding-focused model and a general-purpose one, you get one model that covers both. It sits in the text category and is made by OpenAI.

The model accepts text, images and files as input, and returns text. Its context window is 1,050,000 tokens, with a maximum output of 128,000 tokens. That combination means you can send very large inputs, such as long documents, whole codebases or extended conversation histories, and still have room for a substantial response. Image and file inputs make it possible to mix formats in a single request rather than converting everything to plain text first.

The practical uses follow from those capabilities. Large context plus file input suits document analysis, where you hand over contracts, reports or specifications and ask questions across the whole set. Code review, refactoring and repository-level reasoning are natural fits given the Codex lineage and the ability to hold a large amount of source in context. Image input supports tasks like reading screenshots, diagrams or scanned pages alongside related text. Because the model supports tools and structured outputs, it also works well as the reasoning engine behind agents and pipelines that need to call functions or return data in a fixed schema.

You call it through api14's OpenAI-compatible API using the model id "openai/gpt-5.4". 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. Reasoning and verbosity controls let you trade depth against latency, seed gives you reproducibility where you need it, and tools with tool_choice covers function calling. If you already use an OpenAI client library, pointing it at api14 and changing the model id is enough.

One thing to keep in mind is that output is text only. If your application needs generated images, audio or other output modalities, this model will not cover that part of the pipeline. It is also a frontier model with a very large context window, so for small, simple tasks a lighter model may respond faster and be easier to manage. For long-context reasoning, coding and tool-driven workflows, though, GPT-5.4 is the stronger choice.