Qwen: Qwen3.8 Max (0902)

Qwen3.8 Max (0902) is a text generation model from Qwen, the model family maintained by Alibaba's Qwen team. It is an updated snapshot of Qwen3.8 Max, and the provider describes it as a 2.4-trillion-parameter mixture-of-experts model. The model id on api14 is qwen/qwen3.8-max-0902, and it sits in the text category, which is where you should look for it if you are browsing the catalogue by task type.

The model accepts text, image, and video as input, and it returns text. That combination makes it a multimodal model rather than a text-only one, so you can send a prompt alongside visual material and get a written answer back. The context window is 1,000,000 tokens, which is large enough to hold very long documents, long transcripts, or extended multi-turn conversations without trimming. The maximum output is 131,072 tokens, so responses can be long as well. Reasoning is part of the model's surface area: the supported parameters include reasoning, reasoning_effort, and include_reasoning, which means you can control whether reasoning is exposed and how much effort is spent on it.

Those capabilities point to a few practical uses. Because the input side covers images and video as well as text, the model fits tasks where you need a written answer about visual material, such as describing or summarising what happens in a clip, or answering questions about a set of images. The large context window suits work over long inputs: reviewing lengthy documents, extracting structured information from a big corpus in a single request, or keeping a long conversation coherent. The reasoning controls are useful when a task benefits from step-by-step internal work, and the tool and structured output parameters support workflows where the model has to call functions or return data in a fixed shape.

Calling the model through api14 works the same way as any other OpenAI-compatible endpoint. You send a request to the chat completions route with the model field set to qwen/qwen3.8-max-0902, and you can use the OpenAI client libraries or a plain HTTP request. The request parameters it supports are frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, reasoning_effort, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, and top_p. That covers sampling controls, determinism via seed, stopping conditions, log probability output, structured and JSON-style responses, and tool or function calling. If you need a specific output shape, response_format and structured_outputs are the relevant parameters; if you want the model to use external functions, tools and tool_choice are the ones to set.

One thing to keep in mind is that the model only returns text, so if your application needs generated images, audio, or video as output, this model is not the right fit and you should look for a model whose output modalities match. The very large context window and the 131,072-token output ceiling are generous, but they are still ceilings, and inputs beyond the context length will need to be shortened or split. If your task is purely text in and text out and does not need the multimodal input or the reasoning controls, a smaller text model may be simpler to work with.