Black Forest Labs' precision image generation model with maximum control, reliable text rendering, and complete creative control supporting up to 4MP output
Settled by community votes across 6 shared challenges, with an AI judge weighing in on each.
FLUX.2 [flex]
#14 of 62 in Text-to-Image
Qwen Image Edit 2509
#32 of 32 in Image Editing
Where the votes landed
FLUX.2 [flex]
0%
win rate
Ties
0%
Qwen Image Edit 2509
0%
win rate
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Man and Car in California
Editing“Make a photo of the man driving the car down the California coastline”
AI Judge Analysis
FLUX.2 [flex]
- + Excellent preservation of the car's specific model and geometry.
- + Strong likeness of the man's face and hairstyle from the second source image.
- + Relatively natural-looking hands on the steering wheel.
- − The motion blur on the wheels is slightly inconsistent with the static background elements.
- − The car appears to be right-hand drive based on the man's seating position and steering wheel location, which is inaccurate for a California coast setting.
Qwen Image Edit 2509
- + Beautiful cinematic lighting and high-quality background scenery.
- + Successfully places the man in the driver's seat of a car that maintains the overall look of the source.
- − The man's face loses almost all detailed resemblance to the source image.
- − The perspective of the car has been slightly altered compared to the source image.
- − Notable AI artifacts around the man's hair and the headrest.
Verdict: FLUX.2 [flex] wins because it maintains a far higher degree of subject preservation for both the car and the man. While Qwen Image Edit 2509 creates a more vibrant landscape, it fails to keep the man's specific facial features and introduces more visible artifacts in the interior of the car.
Pose & Character Mashup
Editing“Use Image 1 as the exact pose reference and Image 2 as the character reference. Recreate the person/character from Image 2 in the exact dynamic pose and body position from Image 1. Keep the exact face, hair, clothing style/details, and expression from Image 2. Match the lighting and environment of Image 1. The final image must show the character from Image 2 performing the precise action/pose from Image 1 with perfect anatomy and natural integration.”
AI Judge Analysis
FLUX.2 [flex]
- + Successfully transferred the character's face, sunglasses, scarf, and clothing from Image 2.
- + Accurately recreated the yellow environment and red stool from Image 1.
- + Followed the instruction to put the character from Image 2 into a dynamic pose.
- − Failed to match the 'exact' pose from Image 1, opting for a generic crouch instead of the contorted leg-cross.
- − The anatomy of the feet and the connection to the stool is slightly awkward.
Qwen Image Edit 2509
- + Matches the lighting and background color of the source.
- − Completely failed the character transfer, essentially just providing a low-quality filter of Image 1.
- − Did not incorporate the man, his sunglasses, his scarf, or his clothing from Image 2 as requested.
- − The image quality is significantly degraded compared to the original.
Verdict: FLUX.2 [flex] successfully performed a complex character transfer, bringing the man's identity, accessories, and clothing into the new environment, even if it simplified the requested pose slightly. Qwen Image Edit 2509 failed the prompt entirely, failing to change the character to the one provided in Image 2 and only outputting a lower-resolution version of Image 1. FLUX.2 is the clear winner for actually executing the requested edit.
Bald man challenge
Image Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
FLUX.2 [flex]
- + Excellent preservation of original facial features and lighting.
- + Perfectly follows the request for a realistic, textured, and natural hairline.
- − The hair length is quite short compared to the prompt's request for 'full' hair, appearing more like a buzz cut.
Qwen Image Edit 2509
- + Successfully adds a very full and thick head of hair as requested.
- + Preserves the background and clothing accurately.
- − The hair texture has strange, wavy artifacts that look somewhat synthetic.
- − Slightly alters the shape of the face and forehead compared to the original image.
Verdict: FLUX.2 [flex] produced a much more realistic result that seamlessly integrated with the source image's lighting and texture, although the haircut was shorter than expected. Qwen Image Edit 2509 provided the requested volume, but the hair texture looks artificial and over-processed, and it slightly changed the subject's appearance.
Over-the-top cartoon caricature
Editing“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
AI Judge Analysis
FLUX.2 [flex]
- + Excellent integration of all components, showing multiple dogs in hockey jerseys on a news set.
- + Successfully translates the denim shirt from the source image into the caricature style.
- + High polish and clean cartoon aesthetic that feels cohesive and professional.
- − The caricature of the face is a bit generic and loses some of the specific likeness from the source photo.
- − Text on the news desk ('THEIANAES NEWS') is slightly garbled.
Qwen Image Edit 2509
- + Preserves the selfie-style framing and hand positioning from the source image.
- + Captures the user's specific facial features and dimples better than Model A.
- + Creatively blends the hockey rink with a TV studio background.
- − The 'love for dogs' is represented by a very small, toy-like dog that feels like an afterthought.
- − The large text in the speech bubble is gibberish.
- − The secondary character in the background is not related to the prompt.
Verdict: FLUX.2 [flex] produced a much more finished and thematic image, successfully giving the dogs hockey jerseys and placing the character in a believable newsroom environment. While Qwen Image Edit 2509 did a better job at preserving the specific facial likeness and the selfie perspective, it failed to meaningfully incorporate the dogs and included nonsensical text and characters.
Studio Ghibli Anime Style
Editing“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
AI Judge Analysis
FLUX.2 [flex]
- + Perfectly captures the Studio Ghibli 'watercolor' aesthetic with soft, hand-painted textures.
- + Excellently translates facial expressions into the target art style while keeping the original composition.
- + The color palette is very authentic to the nostalgic Ghibli films like 'Whisper of the Heart'.
- − The woman on the right has lost her original angry expression, appearing oddly content/neutral.
Qwen Image Edit 2509
- + Maintains the original facial expressions much better, especially the angry look of the woman on the right.
- + Preserves the structure of the original image with high fidelity.
- + Good use of warm lighting and blush to enhance the character designs.
- − The style leans more toward modern light-novel or digital illustration rather than the specific 'hand-painted' texture of Studio Ghibli.
- − Visible moiré-like digital patterns on the characters' skin and clothing.
Verdict: FLUX.2 [flex] produced a much more convincing Studio Ghibli imitation, featuring the soft watercolor backgrounds and simple line-work characteristic of the studio's films. While Qwen Image Edit 2509 did a better job of preserving the specific facial expressions (especially the anger of the girlfriend), its overall aesthetic felt like a generic digital filter rather than a hand-painted illustration.
Golden Hour Stroll
Image Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
FLUX.2 [flex]
- + Excellent integration of wind-blown hair that looks natural and realistic.
- + Preserves the identity and features of the woman and dog perfectly.
- + Subtle and effective addition of motion to the dog's tail and the leaves in the foreground.
Qwen Image Edit 2509
- + Successfully adds dynamic leaves and blowing hair.
- + Good preservation of the original scene's lighting and background.
- − The wind-blown hair looks overly stylized/digital with repetitive wave patterns.
- − The added leaves look like flat clip-art stickers layered on top of the image rather than being part of the environment.
- − A strange blurry artifact appeared on the woman's left hand.
Verdict: FLUX.2 [flex] produced a much more realistic and cohesive edit, with hair that flows naturally and motion-blurred leaves that feel like they belong in the 3D space. In contrast, Qwen Image Edit 2509 added hair that looks somewhat artificial and leaves that lack depth, resembling a 2D overlay rather than a photography effect.
Explore each model
Alibaba's Qwen image editing model for instruction-based image modifications and transformations