~ same prompt, different LLM — every model gets the same brief in Blender, you judge the render ~
OpenAI / GPT 6 Astra / thinking:max has entered the arena NEW!
Ramen samurai — printed frameCategory: nprCapture: stillsPrompt: v2Difficulty: ▮▮▮▮▮
TX: ramen-samurai.md — composed prompt
── block: build-request-open@v2 (start, global) ──

Task

Complete the following Blender task:

<task>

One frame from an animated cyberpunk comic that never existed: the Ramen Samurai fighting in a dense multicultural night market, using noodles, broth, steam, bowls, and chopsticks as an improvised martial art. The image should feel like a lost action panel from an aggressively stylized animated graphic novel — playful, kinetic, urban, and unmistakably non-photoreal.

The visual language should lean toward the graphic energy of Into the Spider-Verse without directly copying any single frame: exaggerated perspective, layered ink outlines, halftone shadows, displaced color channels, stepped animation-like poses, screen-tone textures, hand-painted flats, graphic motion shapes, and deliberately imperfect print registration. The final image must look illustrated, printed, and composited from multiple graphic systems rather than rendered with realistic materials.

The setting is a multicultural cyberpunk street district inspired by the lived-in density of Blade Runner and broader global megacity nightlife. Japanese ramen signage shares the frame with Korean hangul, Chinese characters, Arabic storefront lettering, Spanish hand-painted signs, English advertising fragments, South Asian food stalls, improvised electronics, delivery bikes, hanging cables, patched awnings, vending machines, lanterns, and crowded architectural layers. The environment should feel culturally mixed and organically accumulated, not like a generic “Asian neon alley.”

Color palette. Avoid a scene dominated by pink and blue. Use a much wider neon spectrum:

  • acid green and chartreuse signage
  • amber, tungsten orange, and sodium-vapor yellow
  • deep crimson and vermilion lantern light
  • cyan used selectively rather than everywhere
  • violet and magenta only as secondary accents
  • turquoise, lime, dirty white, electric red, and occasional ultraviolet highlights
  • dark indigo, brown-black, and muted concrete tones as grounding colors

The competing colors should overlap through hard-edged graphic light zones rather than smooth realistic illumination. Different signs should create separate color regions across the character and environment. The samurai’s face should remain readable and controlled, with the most intense chromatic misregistration pushed toward the edges of the frame and fast-moving noodle attacks.

The Ramen Samurai. A sharp graphic silhouette: layered street robes mixed with partial samurai armor, practical cyberpunk additions, tied-back hair or topknot, and a sheathed sword still present but clearly secondary. He should not primarily fight with the sword. His entire fighting style is built around ramen preparation transformed into combat.

The noodle fighting style must dominate the image. Long noodles spiral around him in multiple active trajectories, forming whips, loops, lassos, coils, snapping arcs, defensive rings, and striking tendrils. Some noodles wrap around an opponent’s weapon outside the frame; others lash toward silhouettes at the edge of the panel. A broad ribbon of noodles should sweep through the foreground with extreme foreshortening. Chopsticks are used like paired control batons, directing the noodle movement with precise martial gestures.

The noodles should not merely float decoratively. They must visibly perform actions:

  • one strand catches or redirects an incoming blade
  • one loop binds an unseen opponent’s wrist
  • one noodle whip cracks across the foreground
  • one circular coil forms a shield around the samurai
  • several strands trail broth droplets like motion sparks
  • one bundle launches from the bowl in a corkscrew attack
  • loose noodles knot, split, reconnect, and change direction around obstacles
  • noodles wrap around signs, stools, lantern poles, and the ramen cart, making the whole environment part of the choreography

The broth also participates in the combat. It erupts from the bowl in stylized golden-orange arcs, rendered as painted ribbon shapes with halftone droplets and graphic splash bursts. Steam forms bold calligraphic strokes, creature-like curls, and directional glyphs that echo the movement of the noodles. The bowl may rotate or slide across the counter like a spinning weapon platform while remaining readable as a ramen bowl.

The pose should capture a single decisive instant: one hand raised with chopsticks controlling a wide noodle lash, the other hand redirecting a second strand close to the body. His torso twists sharply, robes and armor breaking into angular motion shapes. His eyes lock toward an opponent outside the frame. The body should read clearly even in silhouette, with a strong line of action and no static seated pose.

Composition. Use an extreme comic-book camera: slightly low angle, close enough that the foreground noodle attack crosses the lens, with the samurai placed off-center. Leave open negative space toward the direction of attack so the noodles can travel through the panel. The ramen cart, bowl, signs, and surrounding market should frame the action without competing with the character.

Graphic devices should be built into the scene:

  • thick-to-thin ink outlines
  • coarse halftone shadows
  • offset red, yellow, green, and cyan print layers around moving forms
  • repeated animation-smear noodle silhouettes
  • panel-breaking noodle strands crossing the border
  • angular impact bursts where noodles strike objects
  • small printed sound-effect glyphs integrated into signs and steam
  • screen-tone gradients instead of smooth shading
  • hard two- or three-step lighting
  • selective frame tearing and displaced color blocks
  • motion lines that follow the circular noodle choreography
  • occasional misaligned background duplicates suggesting stepped animation frames

The main image should feel even more NPR than a conventional comic render. Avoid realistic reflections, smooth plastic materials, cinematic depth of field, physically accurate fog, glossy ray-traced surfaces, or ordinary 3D shading. Every visible element should be simplified into bold color groups, drawn contour logic, printed texture, or graphic pattern.

The scene should communicate instantly: this is a Ramen Samurai in a multicultural cyberpunk city, and ramen itself is his weapon, shield, movement system, and fighting discipline.

── block: build-request-close@v2 (end, global) ──
</task>
RUNS ON THIS TASKview in gallery →
RUNPROMPTCONTESTANTSTATUSCOST
r1 · 2026-07-31v2OpenAI / GPT 5.6 Luna / thinking:maxok$0.09
r1 · 2026-07-17v2OpenAI / GPT 5.6 Lunaok$0.04
Votes cast on this bench: 000482 · 157 runs · 25 contestants
Best viewed with Netscape Navigator 4.07 at 800×600 · real runs, live ratings