Qwen: Qwen3.8 27B

Qwen: Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It belongs to the text category on api14, and the provider describes it as suited to coding, professional workflows, research, multimodal interaction, and long-running agent tasks. Because the weights are open, it is also an option for teams that want a Qwen model they can inspect or self-host elsewhere, while still being reachable as a hosted endpoint here.

On input, the model accepts text, images, and video, and it produces text output only. The context window is 1,000,000 tokens, which is large enough to hold very long documents, extended conversation histories, or sizeable codebases in a single request. The maximum output is 235,929 tokens, so responses are not tightly capped. The provider notes flexible thinking, and the model exposes a reasoning parameter alongside include_reasoning and reasoning_effort, which lets you decide whether reasoning is surfaced and how much effort is spent on it.

Those capabilities point to a fairly wide set of practical uses. Coding assistance fits naturally, including work over large repositories where the long context avoids aggressive chunking. Professional and research workflows benefit from being able to pass long reports, papers, or mixed text and image material and ask for a written result. Because video is accepted as input, tasks such as summarizing or querying recorded material are possible. Tool calling and structured outputs make it a reasonable fit for agent-style pipelines that need to run over many steps, and the reasoning controls let you trade depth against latency depending on the task.

You call the model through api14's OpenAI-compatible API using the model id "qwen/qwen3.8-27b". The supported request parameters are frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_a, top_k, top_logprobs, and top_p. In practice this covers sampling controls, deterministic runs via seed, stop sequences, JSON or schema-constrained output through response_format and structured_outputs, and function calling through tools and tool_choice.

One constraint worth noting is that output is text only, so if you need generated images or audio, this model will not provide them. It is a dense 27B model, so a larger model may suit tasks that demand the most difficult reasoning, while this one remains a sensible choice when you want multimodal input, a very long context, and open weights in a single model.