Fast distilled version of Black Forest Labs' FLUX.2 [dev] optimized for speed and cost efficiency.
Settled by community votes across 7 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev] Flash
#5 of 62 in Text-to-Image
Qwen Image Edit 2511
#16 of 32 in Image Editing
Where the votes landed
FLUX.2 [dev] Flash
0%
win rate
Ties
0%
Qwen Image Edit 2511
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 [dev] Flash
- + Excellent preservation of the subject's facial features and distinctive hairstyle.
- + Accurately places the subject inside the specific car model from the source image, including the wooden dashboard and cream interior.
- + Compelling cinematic lighting and motion blur that fits the coastline theme.
- − The steering wheel is positioned too far to the left, appearing to sit in the door frame rather than the dashboard.
Qwen Image Edit 2511
- + Strong source preservation for the clothing, scarf, and hairstyle.
- + The perspective of the road and coastline feels very natural for a California drive.
- + Hand placement on the steering wheel is anatomically plausible.
- − The subject's facial features have been slightly altered, losing the specific likeness of the man in the source photo.
- − The car interior has been generalized and no longer matches the specific Rolls-Royce details from the source.
Verdict: Both models did an excellent job of combining the two source images into a single scene. FLUX.2 [dev] Flash is the winner for its superior ability to maintain the subject's exact likeness and the specific details of the car interior, whereas Qwen Image Edit 2511 slightly altered the man's face and the car's design.
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 [dev] Flash
- + Successfully replicates the yellow background and red ottoman from Image 1.
- + Captures the scarf, sunglasses, and black clothing details from Image 2.
- + Follows the instruction to combine the character features and environment.
- − Fails to replicate the specific pose from Image 1, placing the character in a generic squat/crouch instead.
- − Significant anatomical errors, including a secondary head and arm with long hair emerging from the character's back.
- − Poor feet rendering on the ottoman.
Qwen Image Edit 2511
- + Preserves the character from Image 2 with high fidelity in the background.
- + Retains the original figure from Image 1 in the foreground.
- − Completely fails the core instruction to apply the character from Image 2 to the pose in Image 1, instead showing both people.
- − Ignores the request to match the yellow lighting and environment of Image 1, defaulting to the Image 2 background.
- − The composition is a messy overlay of both images rather than a seamless edit.
Verdict: Both models failed significantly on this complex image editing task. FLUX.2 [dev] Flash attempted to merge the subjects but resulted in a nightmare-like anatomical failure with extra limbs and heads, though it did correctly identify the environment to use. Qwen Image Edit 2511 simply layered the two images on top of each other, completely ignoring the instruction to recreate the character in the specific pose. FLUX.2 [dev] Flash is the technical winner for at least attempting the pose and environment merge, despite the poor execution.
Outfit Transfer Challenge
Editing“Use Image 1 as the base person. Dress them in the exact elaborate outfit from Image 2 (including all layers, accessories, jewelry, and shoes). Carefully adapt the clothing to the body shape and pose in Image 1 while maintaining realistic fabric behavior, correct proportions, and perfect lighting/shadow matching. Keep the person’s exact face, hair, and background completely unchanged.”
AI Judge Analysis
FLUX.2 [dev] Flash
- + Attempts to integrate the clothing while keeping some aspects of the original character's vitiligo.
- − Horrific facial distortion and artifacts.
- − Incorrectly adds heavy jewelry not present in the source outfit.
- − Fails to maintain the person's exact face.
Qwen Image Edit 2511
- + Successfully transfers the entire outfit including pants, shoes, and sunglasses.
- + Maintains a clean, high-quality image with no major artifacts.
- + Matches the pose and anatomy of the person in the background context well.
- − Completely replaces the person's face and skin with the man from Image 2, failing the instruction to keep the base person unchanged.
- − Changes the subject's hair style entirely.
Verdict: Both models failed the core challenge of preserving the base person's identity while changing their clothes. FLUX.2 [dev] Flash produced a nightmarish, distorted facial structure that makes the image unusable, while Qwen Image Edit 2511 simply swapped the person entirely for the man in Image 2. Qwen is the preferred choice only because it produced a coherent, high-quality image, even though it ignored the instruction to keep the base person's face and hair.
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 [dev] Flash
- + Excellent preservation of the original face, lighting, and background.
- + High-density hair consistency that blends well with the existing beard texture.
- − The hair volume is quite extreme, resulting in an 'afro' style that may feel less natural for the character's ethnic features.
- − The hairline is slightly too sharp and straight.
Qwen Image Edit 2511
- + The hair style and texture feel more cohesive and realistic for the man's features.
- + Great preservation of original identity and background elements.
- + Realistic wave and curl texture in the hair.
- − There are slight lighting inconsistencies where the hair meets the forehead.
- − Minimal artifacts around the top edge of the hair against the sky.
Verdict: Both models performed excellently in terms of source preservation, keeping the face, clothing, and background identical to the original image. FLUX.2 [dev] Flash opted for a very high-volume, dense afro-textured head of hair, while Qwen Image Edit 2511 provided a more natural-looking wavy hairstyle that better matched the existing beard and facial structure.
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 [dev] Flash
- + Excellent preservation of the subject's features in a caricature style.
- + Perfectly integrates all requested elements: hockey (dogs in gear, ice rink), anchor job (desk, mic, studio lights), and dogs.
- + High-quality rendering with clean lines and humorous conceptualization.
- − The 'TV SHOW' text is a bit generic.
Qwen Image Edit 2511
- + Captures a humorous, high-energy facial expression suitable for a caricature.
- + Includes the anchor desk and a dog as requested.
- − Completely missed the 'hockey' requirement.
- − Gibberish text in the speech bubble.
- − Anatomical issues with the hands and strange artifacts like the lit candle on the desk.
Verdict: FLUX.2 [dev] Flash is the clear winner as it successfully incorporated every part of the prompt, specifically the hockey element which Qwen Image Edit 2511 ignored. FLUX.2 also did a much better job of maintaining the subject's likeness while translating it into a professional caricature style.
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 [dev] Flash
- + Excellent stylistic adherence to the soft, watercolor-like aesthetic of Studio Ghibli.
- + Creatively transforms the background into a dreamy, flowered landscape while maintaining the core character poses.
- + Soft pastel color palette perfectly matches the requested mood.
- − Significant alterations to the facial structures, making them less recognizable as the individuals from the original meme.
- − Lost the urban setting of the original photo in favor of a field.
Qwen Image Edit 2511
- + High level of source preservation, retaining the original facial features and the urban background elements like the bus.
- + Effectively applies an illustrative style that looks like a high-fidelity digital painting.
- + Accurately captures the lighting effects requested in the prompt without losing structural detail.
- − Style feels more like modern commercial anime than the specific 'hand-painted' Ghibli look requested.
- − The addition of glowing sparkles feels a bit generic compared to the requested Ghibli texture.
Verdict: FLUX.2 [dev] Flash captures the specific 'Ghibli' texture and color palette much better, leaning into a soft watercolor aesthetic, but it changes the scene and faces significantly. Qwen Image Edit 2511 preserves the identity of the original meme perfectly and maintains the city setting, but its art style is more of a generic clean anime filter than the hand-painted, nostalgic look requested.
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 [dev] Flash
- + Successfully added a large volume of flying leaves for a lively feel.
- + The hair motion looks very dynamic and natural.
- + Excellent preservation of the woman's face and original background.
- − Some leaves appear as small, indistinct artifacts or specks rather than clear leaves.
Qwen Image Edit 2511
- + Leaves are rendered with high clarity and saturated colors.
- + Great hair dynamics that follow a consistent wind direction.
- + The dog's tail has a slightly more motion-filled pose compared to the original.
- − Fewer leaves across the composition compared to the other model.
- − Slight alteration to the lighting/shadows on the woman's face.
Verdict: Both models performed excellent edits with strong source preservation. FLUX.2 [dev] Flash captured the 'energetic and lively' feel slightly better by including a greater volume of flying debris, whereas Qwen Image Edit 2511 produced cleaner, more distinct leaf textures but with less overall impact.
Explore each model
Alibaba's Qwen image editing model for instruction-based image modifications and transformations