demobook

OpenArt: Prompt adherence test with Qwen (Quen) Image

Demo summary

The creator demonstrates Alibaba's Qwen image model, which results in a 3D cartoon style when failing to meet the realism requirements of the prompt.

Step-by-step

  1. Select the Qwen image model
  2. Paste your prompt into the input field
  3. Generate the image to check for prompt adherence

Watch out for

  • The model may massively reduce realism and produce 3D cartoonish results if it cannot handle the complexity of the prompt

Highlights

I do have to place this model all the way down in F tier, just because this is by far the [worst]

All demos from “AI Image Generators Ranked from Worst to Best (2026)

  1. 1:420:31Generate images with C-Dream 4.5 and 4The user demonstrates generating a complex scene with a universal prompt using C-Dream 4.5 and C-Dream 4 (BiteDance) within the OpenArt interface.OpenArtText to Image
  2. 5:410:59Testing Flux Pro, Max, and Flex modelsThe video shows a side-by-side comparison of the Flux 2 family (Pro, Max, and Flex) by Black Forest Labs performing the same prompt task.OpenArtAI Image Generator
  3. 7:130:25Artistic image generation with Recraft (Reave)The creator demonstrates the 'Reave' (Recraft) model's ability to generate images with an 80s film aesthetic and high typography quality.RecraftText to Image
  4. 8:060:18Photorealistic generation with OpenArt Photo RealisticThe user demonstrates OpenArt's proprietary model designed to create images that look like raw mobile phone snapshots.OpenArtAI Realistic Image Generator
  5. 10:010:23Historical scene generation with Imagen 4The creator tests Google's Imagen 4 (Image M4) for prompt adherence and historical accuracy in an Aztec-themed generation.OpenArtText to Image
  6. 10:460:20Graphic design generation with Ideogram (Idog) V3The user demonstrates Ideogram V3's performance on a complex prompt, focusing on its smooth aesthetic and typography.IdeogramText to Image
  7. 11:230:29Human realism with Juggernaut FluxThe video showcases Juggernaut Flux (Run Diffusion) generating realistic human skin textures and background text.RunDiffusionAI Realistic Image Generator
  8. 12:080:24Prompt adherence test with Qwen (Quen) ImageCurrentThe creator demonstrates Alibaba's Qwen image model, which results in a 3D cartoon style when failing to meet the realism requirements of the prompt.OpenArtAI Image Generator
  9. Watch “AI Image Generators Ranked from Worst to Best (2026)” →

AI Image Generator

  1. 7:061:43Compare multiple AI models simultaneouslyThe video shows how to use the 'select multiple' feature to compare outputs from GPT Image 2, Nano Banana 2, Recraft, and Kling 3 Omni using the same prompt.Wade McMaster - Creator Impact
  2. 1:322:15Configure SmartShot settings in OpenArt AIThe user demonstrates how to access the SmartShot interface, enter a narrative prompt, select specific character references, and adjust output quality and resolution.Diary Influencer
  3. 8:361:10Generate scene images with character referencesThe demonstration shows how to paste a scene prompt into OpenArt and attach previously generated character images to ensure the correct AI characters appear in the scene.Andrei Ai
  4. 4:550:56Configure image generation settings in OpenArtA walkthrough of the OpenArt creation interface, showing how to select characters, input prompts, use reference images, and set output resolution up to 4K.Isa does AI
  5. 0:190:44Create and manage projects in OpenArtThe user demonstrates how to create a new project named 'fashion portrait' and shows how to rename or delete projects using the workspace panel in OpenArt.OpenArt
  6. 1:240:43Create folders and generate images within themThe user shows how to create a new folder called 'candid shots' and demonstrates that generating images while inside a folder automatically saves them to that specific location.OpenArt
  7. 1:541:20Configuring a Smart Shot prompt in OpenArtThe user demonstrates how to enter a text prompt into the OpenArt Smart Shot interface and configure settings like sheet quality, aspect ratio, and video duration.Ai Lockup
  8. 4:071:27Create a character image with Flux Dev on OpenArtThe user walks through selecting the Flux Dev model in OpenArt, inputting a prompt from ChatGPT, adjusting settings like cinema 16x9 aspect ratio, and upscaling the final result.Dan Kieft
  9. 1:060:42Compare image generation models in OpenArtThe user demonstrates the visual differences between Nano Banana 2, Cedream 4.5, and Cling 3 Omni models using the same text prompt within OpenArt.Roboverse
  10. 3:190:48Save reusable objects and styles in OpenArtThe demonstrator uploads images of a pocket watch, an alleyway, and a specific aesthetic to save them as reusable 'Object', 'Background', and 'Style' assets in the library.Roboverse
  11. 4:301:17Capturing shots with virtual camera controlsThe user shows how to set focal lengths (from 23mm to 300mm) and aspect ratios to 'take a shot' or render a high-fidelity 2D image from a specific perspective within the 3D world.OpenArt
  12. 5:561:17Adjust camera angles with Camera ControlA demonstration of OpenArt's camera angle control tool to change the perspective of a generated image, including low-angle and top-down views.AI Video School