Google: Gemini 2.5 Flash

Google: Gemini 2.5 Flash is a text model from Google, listed on api14 under the model id google/gemini-2.5-flash. It sits in the text category and is described by Google as its state-of-the-art workhorse model, built for advanced reasoning, coding, mathematics and scientific tasks. It includes built-in thinking capabilities, which means the model can work through a problem internally before producing its answer. The provider release date is 2025-06-17, and the knowledge cutoff is 2025-01-31.

The model accepts text, image, audio, video and file inputs, and it returns text only. Its context window is 1,048,576 tokens, which is large enough to hold very long documents, transcripts or codebases in a single request. The maximum output is 65,535 tokens, so long generations are possible within that ceiling. Because input is multimodal, you can send a screenshot, a recording or a PDF alongside your instructions and ask for a written result.

Those capabilities point to a few practical patterns. Long-context work such as summarising large document sets, reviewing lengthy contracts or searching across a repository fits the context window well. Coding and mathematics tasks suit the reasoning focus, especially where the model benefits from thinking before answering. Multimodal inputs open up use cases like describing or extracting information from images, transcribing and analysing audio, or reviewing video content and returning text notes. Structured extraction from mixed media is another option, since the model supports structured outputs.

To call it through api14, use the OpenAI-compatible endpoint with the model id google/gemini-2.5-flash. The supported request parameters are include_reasoning, max_tokens, reasoning, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools and top_p. In practice that means you can control sampling with temperature and top_p, cap output length with max_tokens, set stop sequences, make runs reproducible with seed, request JSON-shaped replies through response_format or structured_outputs, and expose functions to the model with tools and tool_choice. The reasoning and include_reasoning parameters relate to the model's thinking behaviour, so you can influence or surface that process depending on your needs.

One thing to keep in mind is that output is text only, so if you need generated images, audio or video, this model will not produce them. It can read those formats as input but replies in text. If your workload is purely short text exchanges with no need for reasoning, multimodal input or a very large context, a smaller or faster model may be a better fit. For tasks that combine long inputs, mixed media and multi-step reasoning, this model is a reasonable default to evaluate first.