OpenAI: o3 Mini

OpenAI o3 Mini is a text model from OpenAI, released on 31 January 2025. It is positioned as a cost-efficient language model optimized for STEM reasoning tasks, with particular strength in science, mathematics and coding. It sits in the text category and is aimed at developers who need reasoning-heavy answers without reaching for a larger model.

The model accepts text and file inputs and returns text output. It has a context window of 200,000 tokens and a maximum output of 100,000 tokens, so long inputs and long answers are both workable in a single request. Its knowledge cutoff is 31 October 2023. Reasoning is a first-class part of how the model works: the supported parameters include reasoning and include_reasoning, and the provider description notes support for the reasoning_effort parameter, which lets you control how much reasoning effort the model applies to a request.

Those capabilities point to a fairly specific set of use cases. The STEM focus makes it a reasonable fit for solving and explaining mathematics problems, working through science questions, and generating or reviewing code. Because it accepts file inputs alongside text, it can be used in workflows where a document is supplied and the model is asked to reason over its contents. The reasoning controls are useful when you want to trade depth against latency or output length on a per-request basis, and the large context window helps when the reasoning task depends on a lot of surrounding material.

You call the model through api14's OpenAI-compatible API using the model id "openai/o3-mini". The supported request parameters are include_reasoning, max_tokens, reasoning, response_format, seed, structured_outputs, tool_choice and tools. In practice that means you can cap output length with max_tokens, request or expose reasoning traces, constrain the response shape with response_format or structured_outputs, make runs more reproducible with seed, and give the model callable functions through tools and tool_choice.

The main thing to weigh is scope. This is a text-only model: it takes text and file input and produces text, so tasks that need image or audio understanding fall outside it. Its knowledge cutoff of 31 October 2023 means anything after that date has to come from the prompt or from tools rather than from the model itself. If your workload is not reasoning-heavy, or if it depends on modalities this model does not accept, another model is likely a better fit.