Qwen 3.8 27B in the Serverless Cloud API

You can now run Qwen 3.8 27B, Alibaba's open-weights vision language model, in Roboflow's Serverless Cloud API. Send an image and a question, such as "Is this dump truck loaded?", and it answers in text. It works with short-answer tasks including visual question answering, single-label and multi-label classification, and more. Because it runs on Roboflow's GPUs, no outside provider account is required.

If you want to use it, post to https://serverless.roboflow.com/infer/lmm/qwen3_8-27b with your Roboflow API key, a model_id of qwen3_8-27b, your image, and your prompt. The answer comes back as text in the response field.