The Qwen-Edit images from my GenAI Image Editing Showdown site were all generated from a ComfyUI workflow on my machine - it's shockingly good for an open-weight model. It was also the only model that scored a passing grade on the Van Halen M&M test (even compared against Nanobanana)
Ha I created a Van Halen M&M test for text prompts. I would include an instruction demanding that the response contain <yellow_m&m> and <red_m&m> but never <brown_m&m>. Then I would fail any llm that did not include any m&ms, or if they wrote anything about the <brown_m&m> in the final output.
https://genai-showdown.specr.net/image-editing