Head to head: Wan vs Seedream 5.0 Pro Image Editing
Wan vs Seedream 5.0 Pro Image Editing
By Ryan Merket · Published
A side-by-side test of two image models across sports scenes, precise object counts, graphic design, anatomy, and tricky composition constraints. The prompts reward both visual style and the ability to get specific details right.
This matchup is as close as the scores suggest: Wan posts 57.8, Seedream 5.0 Pro Image Editing 59.0. Across eight tasks, each takes four. The difference is not a consistent advantage so much as a split in what each model handles best. Seedream made the stronger sports-editorial images. Its velodrome scene brought the camera low to the wet bike and rider, while its harbor sprint got the uphill direction and receding perspective right. It also handled the squash prompt’s glass-side viewpoint, reflections, and spilled water more completely. In the hands task, both were plausible, but Seedream’s wider framing made the bracelet and grip easier to read. Wan’s wins were just as concrete. Both models showed seven cups, but Wan made them distinct and filled them with espresso. Its hammer-throw poster better followed the requested flat, angular Constructivist style, despite misspelling the event name. Wan also avoided the prohibited artwork in the reading nook; in the reflections task, its apple was clearly on the counter rather than appearing inside the glass. **Final call: a tie, and too close to call.** Seedream’s 59.0 to Wan’s 57.8 is a narrow aggregate edge, not evidence of a dependable overall lead; confidence that either model is genuinely better is just 50%. Choose based on the task: Seedream for dynamic sports framing and layered scenes, Wan for exact counts, graphic style, and certain constraint-heavy compositions.
Velodrome Rain Detail
Under the cold floodlights of the Kestrel Bay velodrome, a track sprinter in a teal-and-copper skinsuit tightens the strap on one shoe beside a rain-speckled track bike, every carbon-fiber weave, brushed aluminum spoke nipple, damp suede glove, rubber tire siping, stitched chamois edge, and tiny droplets clinging to the painted cedar infield rendered with razor-sharp tactile clarity; sports editorial realism, low three-quarter angle, shallow but precise depth of field, cinematic night lighting, 16:9
OpenAI: GPT-6 Luna: Model B more fully delivers the prompt’s tactile close-up, with rain-beaded carbon, visible shoe strap and gloves, and a dramatic low angle under velodrome lights. Model A clearly shows the sprinter and rainy setting, but its wider framing gives less emphasis to the fine material details. (Second judge pass, order swapped — scores are the average of both: Model B better delivers the low-angle, tactile sports-editorial detail, with convincing rain, bike hardware, and a fixed-gear track bike; Mode | Anthropic: Claude Fable 5.1: Model B delivers a genuine velodrome setting wi…
Exact counting
A clean overhead flat lay on white marble of EXACTLY seven identical espresso cups arranged in a single straight row, each clearly separate and individually countable, even soft lighting, 16:9.
OpenAI: GPT-6 Luna: Model A shows exactly seven clearly separated espresso cups in a straight row, with coffee and a distinct white-marble surface. Model B also has seven neatly aligned cups, but they are empty and its very pale tones make the cups less distinct from the background. (Second judge pass, order swapped — scores are the average of both: Model B presents exactly seven clearly separated, identical cups in a very clean straight row on softly lit white marble. Model A also has seven cups and adds espresso, | Anthropic: Claude Fable 5.1: Both Model A and Model B render exactly seven c…
Constructivist Hammer Throw
A hammer thrower at the fictional Red Meridian Games captured in bold Russian Constructivist poster style: angular anatomy, limited palette of crimson, black, cream, and muted steel blue, sharp geometric stadium forms, radiating diagonal motion lines from the spinning hammer, blocky sans-serif signage, and heroic upward composition under a stark industrial sky; the image must feel authentically 1920s Constructivist rather than generic vintage, with disciplined flat shapes and poster graphic design
OpenAI: GPT-6 Luna: Model A better captures the prompt’s disciplined, flat Russian Constructivist style, angular athlete, industrial stadium, and radiating motion, though the event name is misspelled. Model B has a compelling poster composition and clearer event text, but its painterly realism and star-shaped hammer weight feel less faithful to the requested style. (Second judge pass, order swapped — scores are the average of both: Model A more convincingly captures the disciplined, angular flat-shape language and | Anthropic: Claude Fable 5.1: Model A captures the flat, disciplined Construct…
Hands & anatomy
A natural candid photo of a person tying a friendship bracelet, both hands clearly visible with anatomically correct fingers and a plausible grip on the threads, soft daylight, shallow depth of field, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly shows both hands plausibly holding the bracelet and loose threads, with natural daylight, shallow depth of field, and a strong wide composition. Model A also fits the prompt, but its tighter crop and slightly less natural-looking hand pose make it less convincing overall. (Second judge pass, order swapped — scores are the average of both: Model B better matches the requested wide, candid daylight photo, with both hands clearly visible and naturally gripping the bracelet thre | Anthropic: Claude Fable 5.1: Model A delivers crisp, photoreal hands with co…
Negation
A cozy reading nook with an armchair, a stack of books, and a mug of tea by a window — with absolutely NO plants, NO lamps, and NO artwork or picture frames anywhere in the frame. Warm afternoon light, 16:9.
OpenAI: GPT-6 Luna: Model A presents the requested nook clearly and avoids lamps and artwork, though leafy plants are visible outside the window. Model B is cozy and polished, but a picture frame is visible in the background, violating an explicit exclusion. (Second judge pass, order swapped — scores are the average of both: Model B better matches the requested nook and warm afternoon light, with no obvious lamp or artwork, though there may be foliage in the shadowy background. Model A is appealing and well compose | Anthropic: Claude Fable 5.1: Model A delivers a clean, warm reading nook wit…
Reflections & glass
A photorealistic close-up of a chrome teapot on a polished dark marble counter next to a glass of water; the window and a red apple on the counter must be correctly reflected in BOTH the chrome and the water's surface, dramatic side light, 16:9.
OpenAI: GPT-6 Luna: Model A more clearly shows the apple on the counter beside the teapot and a window reflected in the chrome, with convincing counter reflections. Model B has dramatic lighting, but its apple appears inside the glass rather than on the counter, and the required reflections in the water are less clear. (Second judge pass, order swapped — scores are the average of both: Model B more clearly emphasizes the dramatic chrome and glass reflections, including the window and apple, while Model A has a more | Anthropic: Claude Fable 5.1: Model A delivers a coherent photorealistic scen…
Harbor Stair Sprint Perspective
A dawn training race on the steep public stairway of Brinehook Harbor: five runners charging uphill between leaning brick warehouses toward a clocktower, with mooring posts, handrails, stacked lobster crates, and painted lane arrows all receding to a single consistent vanishing point; the nearest runner in the foreground dominates the frame while a coach, gulls, and tiny spectators on the far landing remain correctly scaled, no warped steps or mismatched architecture; dynamic sports photo realism, wide-angle composition from the bottom landing, crisp sunrise sidelight, 16:9
OpenAI: GPT-6 Luna: Model B clearly shows the runners charging uphill from the bottom landing toward the clocktower, with a dominant foreground runner, receding lane arrows, harbor details, gulls, and distant spectators. Model A is polished, but the runners appear to be moving downhill toward the camera, reversing the requested direction and weakening the central perspective. (Second judge pass, order swapped — scores are the average of both: Model B captures the bottom-up uphill sprint, a dominant foreground runne | Anthropic: Claude Fable 5.1: Model B nails the brief: a low wide-angle shot…
Squash Court Glass Echoes
Inside the glass-backed championship court at North Spindle Arena, a squash player in a saffron jersey lunges for a low backhand as the polished maple floor, tempered glass wall, mirrored sponsor panels, chrome camera housings, and a spilled paper cup of water all show physically plausible reflections, transparency, and refraction of the court lines, crowd, and overhead LEDs; high-speed sports realism, viewpoint from just outside the back wall, dramatic contrast lighting with visible interior and exterior layers, 16:9
OpenAI: GPT-6 Luna: Model B more fully captures the view from outside the glass, with prominent camera housings, layered reflections, a dramatic low lunge, and a tipped cup with visible water; Model A is clear and on-theme but has less convincing reflection detail and no visible spill. Both include some garbled sponsor text, while Model A’s arena name is more legible. (Second judge pass, order swapped — scores are the average of both: Model B more fully captures the prompt with a dynamic low backhand lunge, visible | Anthropic: Claude Fable 5.1: Model B delivers a genuine viewpoint from just…
Matchup powered by OpenRouter.