Tasks and inputs
Qwen Image 2.1 PE I2I is listed for Text generation, Prompt Enhancement. It accepts text, image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Oct 8, 2026
Qwen Image 2.1 PE I2I rewrites image editing instructions with image references into detailed prompts for a downstream image model.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Qwen Image 2.1 PE I2I rewrites image editing instructions with image references into detailed prompts for a downstream image model.
This is a prompt-rewriting checkpoint based on a fine-tuned Qwen3.5-VL 9B model. It reads an editing instruction and input images, then turns the request into a more precise editing prompt. The output is text and structured information, not the finished image.
Specify which details should change and which must remain. The model can refer to multiple supplied images by numbered image references, so keep their order consistent. Review the rewritten instructions before passing them to an editor.
The model card documents a JSON result after a reasoning block. Parse and validate the expected fields instead of forwarding the entire raw response as an image prompt. Keep the original brief, rewritten result and generated image together so changes can be traced during evaluation.
The weights use the Qwen Research License Agreement, which differs from permissive terms on some other Qwen models. Follow the model’s processor and template. This prompt enhancer has its own memory footprint and latency in addition to the downstream image model, and no ordinary image-generation token price should be inferred from the checkpoint.
Compare the model on the work you need to complete.
Qwen Image 2.1 PE I2I is listed for Text generation, Prompt Enhancement. It accepts text, image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
Documented limits: 262.14K context. Measure a completed task using the prompts, documents and tools your application will use. Include reasoning and retries in the result.
The weights are downloadable under Qwen Research License Agreement. Check runtime support and size the complete model and working context for your expected concurrency.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
Qwen
Qwen Image 2.1 PE T2I rewrites short image briefs into detailed prompts for a downstream image model.