# Qwen Image 2.1 API via PiAPI > Qwen Image 2.1 Turbo generates one image from a text prompt. Use 8 steps by default, with 1–16 accepted; each side is 64–4096 pixels and the total area is at most 1,763,584 pixels. Landing page: https://piapi.ai/qwen-image-2-1 Base URL: https://api.piapi.ai Authentication: use the API key from your PiAPI workspace. Documentation: https://piapi.ai/docs/qwen-image-api/text-to-image ## Product details - Category: image - Vendor: alibaba ## Runtime targets ### Qwen Image 2.1 Turbo / Text to Image POST https://api.piapi.ai/api/v1/task ```json { "model": "Qubico/qwen-image-2.1", "task_type": "txt2img", "input": { "prompt": "An editorial photograph of a ceramic teapot on a wooden table in morning light.", "width": 1024, "height": 1024, "steps": 8, "seed": -1, "negative_prompt": "", "flow_shift": 1 } } ``` GET https://api.piapi.ai/api/v1/task/{task_id} Output fields: image_url, image_urls Workspace: https://piapi.ai/workspace/qwen-image-2-1?operation=txt2img Documentation: https://piapi.ai/docs/qwen-image-api/text-to-image - default: USD 0.015 / image ### Related entities - Qwen Image 2.1 API - Qwen Image 2.1 Turbo - Alibaba - PiAPI ## FAQ ### Does Qwen Image 2.1 support image editing? This API supports text-to-image only. Existing Qwen Image provides a separate image-edit operation. ### Which step count should I use? Use the default and recommended 8 steps. Custom schedules at 1–7 or 9–16 steps may reduce quality; more steps are not guaranteed to improve results. ### What dimensions are supported? Width and height default to 1024 and each accepts 64–4096 pixels. Their product must not exceed 1,763,584 pixels: 1328 × 1328 and 4096 × 64 fit, while 4096 × 4096 does not. ### Do negative_prompt and flow_shift change the result? No. Both fields are accepted for compatibility but ignored by Turbo at every supported step count. ### How much does generation cost? One image costs $0.015. Create a task, then poll its task_id until completed to retrieve the output.