Approved models
86
Records published from this collection
Provider information
lab
Review approach
Verify each model’s source, terms, capabilities, and current availability
Qwen publishes language and multimodal model families. This directory separates the capability of a named checkpoint from the services and terms of an inference host.
Qwen Image 2.1 combines image generation, editing and native transparent RGBA output. Its documented reference-image workflow makes it relevant for product assets and iterative edits.
Give a visual brief with the exact text, composition and parts that must remain unchanged. For an edit, compare subject identity and layout as well as the requested change. For a transparent asset, inspect the alpha edge against both light and dark backgrounds.
The model card names the pipeline and current dependency requirements. A parameter count for the visual-generation component does not describe the full pipeline’s memory use. Test the chosen dimensions, precision and reference-image count on your deployment.
Qwen Image 2.1 has Qwen Research licensing, rather than an assumed Apache licence. Read the checkpoint’s terms before choosing it for a project. A managed API also has its own usage limits and billing units.
For a hosted alternative, see the Qwen versus Nano Banana guide. Measure assets accepted against the same brief, the number of revisions and the time needed to finish each output.
Review each record individually; collection membership does not imply a recommendation or guarantee of availability.
1–60 of 86 models · select a column to sort
About this order: Recommended shows our editorial shortlist first, followed by recent releases. The spec score is a 0–100 figure SyncDev calculates from each vendor's published specifications — context window, reasoning support, input modalities, tool calling, maximum output and how recently the model shipped. It measures breadth of capability, not benchmark performance, so a higher-scoring model is not automatically the better choice for your task.