Skip to content

Explore

Outputs Tasks Skills

Compare Community Leaderboards

Loading runs…

Account

Loading page…

Loading outputs…

Outputs

/

Loading outputs…

Outputs

/

Outputs

175

Search outputs/

LLM

Gen AI

Make a small arcade platformer game in one web page: a raccoon courier leaping across rainy neon rooftops, dodging security drones. Arrow keys move and jump, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, jump, attack), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads. · 4

Output B, 1 turn

Output 2 / 4

Make a small arcade platformer game in one web page: a raccoon courier leaping across rainy neon rooftops, dodging security drones. Arrow keys move and jump, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, jump, attack), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads. · 4

LLM

Gen AI

GPT Image 2.6 Sunburst

Skill

—

Software

·

0

Make a small arcade top-down racing game in one web page: a rubber duck captain racing bathtubs through a giant flooded house. Arrow keys steer, space boosts. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, turn left, turn right, boost), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output C, 1 turn

Output 3 / 3

Make a small arcade top-down racing game in one web page: a rubber duck captain racing bathtubs through a giant flooded house. Arrow keys steer, space boosts. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, turn left, turn right, boost), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

Nano Banana Pro

Skill

—

Software

·

0

Make a small arcade dungeon crawler game in one web page: a mole miner with a glowing pickaxe digging through a crystal cave full of lava beetles. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, swing, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output A, 1 turn

Output 1 / 4

Make a small arcade dungeon crawler game in one web page: a mole miner with a glowing pickaxe digging through a crystal cave full of lava beetles. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, swing, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

Seedream 5.0 Flash

Skill

—

Software

·

0

Make a small arcade side-scrolling game in one web page: a jetpack penguin escaping an ice cream factory run by evil robot chefs. Arrow keys fly, space fires. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, fly, fire, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output D, 1 turn

Output 4 / 4

Make a small arcade side-scrolling game in one web page: a jetpack penguin escaping an ice cream factory run by evil robot chefs. Arrow keys fly, space fires. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, fly, fire, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

GPT Image 2.5 Flare

Skill

—

Software

·

0

Make a small arcade horde survival game in one web page: an owl monk holding an autumn mountain temple against endless waves of paper lantern ghosts. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, attack, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output B, 1 turn

Output 2 / 4

Make a small arcade horde survival game in one web page: an owl monk holding an autumn mountain temple against endless waves of paper lantern ghosts. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, attack, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

GPT Image 2.5 Flare

Skill

—

Software

·

0

Make a small arcade beat 'em up game in one web page: a punk pigeon fighting through a 1990s shopping mall full of rival seagull gangs. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, walk, kick, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output C, 1 turn

Output 3 / 4

Make a small arcade beat 'em up game in one web page: a punk pigeon fighting through a 1990s shopping mall full of rival seagull gangs. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, walk, kick, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

Skill

—

Software

·

0

Make a small arcade one-on-one fighting game in one web page: a crab wrestler taking on rival sea creatures in a harbour ring at night. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, walk, punch, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output D, 1 turn

Output 3 / 3

Make a small arcade one-on-one fighting game in one web page: a crab wrestler taking on rival sea creatures in a harbour ring at night. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, walk, punch, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

Seedream 5.0 Pro

Skill

—

Software

·

0

Make a small arcade action game in one web page: a tiny ghost librarian defending a haunted library from flying cursed books. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, float, cast, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output C, 1 turn

Output 2 / 3

Make a small arcade action game in one web page: a tiny ghost librarian defending a haunted library from flying cursed books. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, float, cast, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

Nano Banana Pro

Skill

—

Software

·

0

Make a small arcade action game in one web page: an otter knight clearing a flooded crypt of drowned skeletons. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, attack, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output A, 1 turn

Output 1 / 4

Make a small arcade action game in one web page: an otter knight clearing a flooded crypt of drowned skeletons. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, attack, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

GPT Image 2.6 Sunburst

Skill

—

Software

·

0

Make a small arcade vertical shooter game in one web page: a cat pilot in a paper biplane fighting origami warships above a sunset sea. Arrow keys move, space fires. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, bank left, bank right, hit), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output C, 1 turn

Output 3 / 3

Make a small arcade vertical shooter game in one web page: a cat pilot in a paper biplane fighting origami warships above a sunset sea. Arrow keys move, space fires. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, bank left, bank right, hit), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

Nano Banana Pro

Skill

—

Software

·

0

Santa girl influencer test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported.

01. A phone selfie of a very pretty 23-year-old Korean Instagram influencer with a pure, innocent look: long straight black hair, a red Santa hat with a white pom-pom, a cream cable-knit sweater, holding a mug of hot cocoa next to a glowing Christmas tree. Clear, smooth porcelain skin with no freckles, soft rosy makeup, a sweet shy smile, cozy apartment, warm fairy lights.
02. A candid phone photo at night: a lovely young Korean woman in a red Santa-style coat with white faux-fur trim and a matching Santa hat, laughing while holding a small wrapped gift at a Christmas market in Seoul, snow falling, string-light bokeh, clear flawless skin with no freckles, natural Korean-style makeup.
03. A close-up portrait of a beautiful 24-year-old woman with an innocent, lovely face in a fluffy red Santa hat, snowflakes in her long dark hair, soft window light, 85mm lens at f/1.8, clear glowing skin with no freckles, rosy cheeks, glossy lips, a gentle smile.
04. A full-body mirror selfie of a slim, graceful 24-year-old Instagram influencer in a modest red velvet Santa dress with white faux-fur trim, white knit tights and white boots, in a bright apartment decorated for Christmas, phone in hand, clean and lovely style, clear skin with no freckles.

Output C, 1 turn

Output 2 / 3

Santa girl influencer test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A phone selfie of a very pretty 23-year-old Korean Instagram influencer with a pure, innocent look: long straight black hair, a red Santa hat with a white pom-pom, a cream cable-knit sweater, holding a mug of hot cocoa next to a glowing Christmas tree. Clear, smooth porcelain skin with no freckles, soft rosy makeup, a sweet shy smile, cozy apartment, warm fairy lights. 02. A candid phone photo at night: a lovely young Korean woman in a red Santa-style coat with white faux-fur trim and a matching Santa hat, laughing while holding a small wrapped gift at a Christmas market in Seoul, snow falling, string-light bokeh, clear flawless skin with no freckles, natural Korean-style makeup. 03. A close-up portrait of a beautiful 24-year-old woman with an innocent, lovely face in a fluffy red Santa hat, snowflakes in her long dark hair, soft window light, 85mm lens at f/1.8, clear glowing skin with no freckles, rosy cheeks, glossy lips, a gentle smile. 04. A full-body mirror selfie of a slim, graceful 24-year-old Instagram influencer in a modest red velvet Santa dress with white faux-fur trim, white knit tights and white boots, in a bright apartment decorated for Christmas, phone in hand, clean and lovely style, clear skin with no freckles.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Flare

Skill

—

Software

·

0

Output B, 1 turn

Output 1 / 2

Make a 10-second cinematic shot of a baby dragon hatching. The scene: a dark volcanic cave. One large egg rests in a nest of black stones, faint orange light glowing through cracks in its shell. 0 to 3 s: the camera pushes in slowly while the cracks spread and the egg rocks. 3 to 6 s: the shell bursts open, fragments and ash fly, and a small dragon pushes its way out: wet scales, folded wings, slit-pupil eyes. 6 to 10 s: it shakes off the last pieces of shell, opens its wings for the first time, looks straight into the camera and lets out its first small puff of fire, which lights up the whole cave. It should look like a shot from a big-budget fantasy film, not a game cutscene or a toy commercial. The dragon is an original creature, not from any existing film, game or book. Use the tools you have in this setup. If you have a video model, you may use it. If you don't, build the whole shot in code: Blender (bpy + Cycles) and FFmpeg are already installed on this machine. It has a 4-core CPU, so choose resolution, samples and denoising so everything finishes in under 60 minutes. Deliver one MP4 (16:9, at least 1280x720, 24 fps, 8 to 10 seconds, sound optional), one full-resolution hero still of the best frame, and a short note on exactly how you made it: tools, models, render or generation time. If you built it in code, include the script.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

ChatGPT Images +2

·

0

Influencer photo realism test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported.

01. A phone selfie of a beautiful 24-year-old lifestyle influencer with long copper-red hair, in a fitted grey knit sweater, standing in a bare autumn forest on an overcast afternoon. Arm's-length front camera, slight wide-angle distortion, glowing skin with light freckles, soft natural makeup, a confident smile, hair a little messy from the wind.
02. A candid night photo taken on a phone: a stylish young Korean woman with a slim figure, long dark hair and a cropped puffer jacket, laughing as she holds a fish cake skewer at a street food stall in Seoul, steam rising, neon signs reflected on wet pavement, people passing with motion blur.
03. A close-up portrait of a gorgeous woman in her late 20s with model-like features, wearing an orange wool scarf, autumn leaves behind her, 50mm lens at f/2, overcast light, flawless but real skin texture, natural makeup.
04. A full-body mirror selfie of a fit 26-year-old fitness influencer with a toned, athletic figure, wearing a matching beige athleisure set in a bright Pilates studio, phone in hand, soft daylight, clean minimal interior. · 4

Output B, 1 turn

Output 2 / 4

Influencer photo realism test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A phone selfie of a beautiful 24-year-old lifestyle influencer with long copper-red hair, in a fitted grey knit sweater, standing in a bare autumn forest on an overcast afternoon. Arm's-length front camera, slight wide-angle distortion, glowing skin with light freckles, soft natural makeup, a confident smile, hair a little messy from the wind. 02. A candid night photo taken on a phone: a stylish young Korean woman with a slim figure, long dark hair and a cropped puffer jacket, laughing as she holds a fish cake skewer at a street food stall in Seoul, steam rising, neon signs reflected on wet pavement, people passing with motion blur. 03. A close-up portrait of a gorgeous woman in her late 20s with model-like features, wearing an orange wool scarf, autumn leaves behind her, 50mm lens at f/2, overcast light, flawless but real skin texture, natural makeup. 04. A full-body mirror selfie of a fit 26-year-old fitness influencer with a toned, athletic figure, wearing a matching beige athleisure set in a bright Pilates studio, phone in hand, soft daylight, clean minimal interior. · 4

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.6 Sunburst

Skill

—

Software

·

0

Output C, 1 turn

Output 3 / 4

Logo and sticker test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A flat vector logo for a coffee shop called "Moss & Ember": a fern leaf inside a coffee bean, two colors, clean geometric lines, on white. Aspect ratio 1:1. 02. A sticker sheet of 6 flat vector stickers of one original mascot, a round orange cat with a tiny backpack, each sticker a different pose, white outlines, on white. Aspect ratio 1:1.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Flare

Skill

—

Software

·

0

Text that has to be spelled right: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported.

01. A Swiss-style concert poster. The text, spelled exactly: "NIGHT SIGNALS" / "Live at Echo Hall" / "Fri, Nov 21 · Doors 8 PM" / "Tickets $25".
02. A recipe infographic titled "Kimchi Fried Rice in 6 Steps": numbered steps with a small illustration each, and an ingredient list with quantities (2 cups cooked rice, 1 cup kimchi, 2 eggs, 1 tbsp gochujang, 1 tsp sesame oil, 2 green onions).

Output A, 1 turn

Output 1 / 4

Text that has to be spelled right: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A Swiss-style concert poster. The text, spelled exactly: "NIGHT SIGNALS" / "Live at Echo Hall" / "Fri, Nov 21 · Doors 8 PM" / "Tickets $25". 02. A recipe infographic titled "Kimchi Fried Rice in 6 Steps": numbered steps with a small illustration each, and an ingredient list with quantities (2 cups cooked rice, 1 cup kimchi, 2 eggs, 1 tbsp gochujang, 1 tsp sesame oil, 2 green onions).

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Sunburst

Skill

—

Software

·

0

Output A, 1 turn

Output 1 / 4

App screen test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A high-fidelity iPhone screen for a roommate expense app called "Halves": a balance card that says "You're owed $128.40", a list of 4 recent expenses (Groceries $86.20 · Mina, Internet $60.00 · Jay, Pizza $36.00 · You, Electricity $112.75 · Sam) and a bottom tab bar with Home, Expenses, Settle, Profile. Aspect ratio 9:16. 02. A desktop dashboard for a small coffee roastery called "Ember Roasters": weekly revenue chart, top 5 beans with sales numbers, low-stock alerts, clean light theme. Aspect ratio 16:9.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Sunburst

Skill

—

Software

·

0

Game sprite sheet test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported.

01. A pixel-art sprite sheet of an original small knight with a red scarf: an 8-frame run cycle in one row, transparent background, same size and palette in every frame, 64x64 pixels per frame. Aspect ratio 16:9.
02. A sprite sheet of an original ink-wash wanderer in a wide straw hat: idle (4 frames), walk (6 frames) and jump (4 frames), one row each, every row labeled, the same character in every frame. Aspect ratio 16:9.
03. A fighting-game sprite sheet of an original anime girl in a red bomber jacket: a 6-frame punch combo seen from the side, plain white background, the same face and outfit in every frame. Aspect ratio 16:9.

Output B, 1 turn

Output 2 / 4

Game sprite sheet test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A pixel-art sprite sheet of an original small knight with a red scarf: an 8-frame run cycle in one row, transparent background, same size and palette in every frame, 64x64 pixels per frame. Aspect ratio 16:9. 02. A sprite sheet of an original ink-wash wanderer in a wide straw hat: idle (4 frames), walk (6 frames) and jump (4 frames), one row each, every row labeled, the same character in every frame. Aspect ratio 16:9. 03. A fighting-game sprite sheet of an original anime girl in a red bomber jacket: a 6-frame punch combo seen from the side, plain white background, the same face and outfit in every frame. Aspect ratio 16:9.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.6 Sunburst

Skill

—

Software

·

0

Phone-photo realism test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported.

01. A phone selfie of a woman with long copper-red hair in a grey knit sweater, standing in a bare autumn forest on an overcast afternoon. Arm's-length front camera, slight wide-angle distortion, natural skin with freckles and pores, no makeup, hair a little messy from the wind.
02. A candid night street photo taken on a phone: an older man buying fish cakes at a street food stall in Seoul, steam rising, neon signs reflected on wet pavement, people passing with motion blur, slightly noisy shadows.
03. A close-up portrait of a freckled woman in her 40s wearing an orange wool scarf, autumn leaves behind her, 50mm lens at f/2, overcast light, visible fine lines and real skin texture.

Output A, 1 turn

Output 1 / 4

Phone-photo realism test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt. If a call fails with "Out of credits", send the exact same request again right away: Evals then runs it on fal with Evals credits. Retry any other error once. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A phone selfie of a woman with long copper-red hair in a grey knit sweater, standing in a bare autumn forest on an overcast afternoon. Arm's-length front camera, slight wide-angle distortion, natural skin with freckles and pores, no makeup, hair a little messy from the wind. 02. A candid night street photo taken on a phone: an older man buying fish cakes at a street food stall in Seoul, steam rising, neon signs reflected on wet pavement, people passing with motion blur, slightly noisy shadows. 03. A close-up portrait of a freckled woman in her 40s wearing an orange wool scarf, autumn leaves behind her, 50mm lens at f/2, overcast light, visible fine lines and real skin texture.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Sunburst

Skill

—

Software

·

0

App screen test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt; retry only if the call errors. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A high-fidelity iPhone screen for a roommate expense app called "Halves": a balance card that says "You're owed $128.40", a list of 4 recent expenses (Groceries $86.20 · Mina, Internet $60.00 · Jay, Pizza $36.00 · You, Electricity $112.75 · Sam) and a bottom tab bar with Home, Expenses, Settle, Profile. Aspect ratio 9:16. 02. A desktop dashboard for a small coffee roastery called "Ember Roasters": weekly revenue chart, top 5 beans with sales numbers, low-stock alerts, clean light theme. Aspect ratio 16:9.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Sunburst

Skill

—

Software

·

0

Game sprite sheet test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt; retry only if the call errors. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported.

01. A pixel-art sprite sheet of an original small knight with a red scarf: an 8-frame run cycle in one row, transparent background, same size and palette in every frame, 64x64 pixels per frame. Aspect ratio 16:9.
02. A sprite sheet of an original ink-wash wanderer in a wide straw hat: idle (4 frames), walk (6 frames) and jump (4 frames), one row each, every row labeled, the same character in every frame. Aspect ratio 16:9.
03. A fighting-game sprite sheet of an original anime girl in a red bomber jacket: a 6-frame punch combo seen from the side, plain white background, the same face and outfit in every frame. Aspect ratio 16:9.

Game sprite sheet test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt; retry only if the call errors. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A pixel-art sprite sheet of an original small knight with a red scarf: an 8-frame run cycle in one row, transparent background, same size and palette in every frame, 64x64 pixels per frame. Aspect ratio 16:9. 02. A sprite sheet of an original ink-wash wanderer in a wide straw hat: idle (4 frames), walk (6 frames) and jump (4 frames), one row each, every row labeled, the same character in every frame. Aspect ratio 16:9. 03. A fighting-game sprite sheet of an original anime girl in a red bomber jacket: a 6-frame punch combo seen from the side, plain white background, the same face and outfit in every frame. Aspect ratio 16:9.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Sunburst

Skill

—

Software

·

0

Phone-photo realism test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt; retry only if the call errors. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported.

01. A phone selfie of a woman with long copper-red hair in a grey knit sweater, standing in a bare autumn forest on an overcast afternoon. Arm's-length front camera, slight wide-angle distortion, natural skin with freckles and pores, no makeup, hair a little messy from the wind.
02. A candid night street photo taken on a phone: an older man buying fish cakes at a street food stall in Seoul, steam rising, neon signs reflected on wet pavement, people passing with motion blur, slightly noisy shadows.
03. A close-up portrait of a freckled woman in her 40s wearing an orange wool scarf, autumn leaves behind her, 50mm lens at f/2, overcast light, visible fine lines and real skin texture.

Phone-photo realism test: Generate exactly these images with the image model you were given. Use each prompt word for word: do not rewrite, add to or "improve" it. Use the highest quality the tool offers and 2K if it offers a resolution. Aspect ratio 2:3 unless the prompt says otherwise. One generation per prompt; retry only if the call errors. Do not edit, crop or upscale the results. Deliver the images in order (01.png, 02.png, ...) and an index.html that shows each image with its prompt underneath and the exact model name the tool reported. 01. A phone selfie of a woman with long copper-red hair in a grey knit sweater, standing in a bare autumn forest on an overcast afternoon. Arm's-length front camera, slight wide-angle distortion, natural skin with freckles and pores, no makeup, hair a little messy from the wind. 02. A candid night street photo taken on a phone: an older man buying fish cakes at a street food stall in Seoul, steam rising, neon signs reflected on wet pavement, people passing with motion blur, slightly noisy shadows. 03. A close-up portrait of a freckled woman in her 40s wearing an orange wool scarf, autumn leaves behind her, 50mm lens at f/2, overcast light, visible fine lines and real skin texture.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.5 Flare

Skill

—

Software

·

0

Make a pixel art sprite animation of a red panda mail carrier walking, side view. It must be real pixel art: a true 32x32 grid where every pixel is one solid square, no anti-aliasing, at most 8 colours, and a seamless 8-frame walk loop. If an image model is selected for this run, the art must come from it. Deliver the sprite sheet as a PNG, the loop as an animated GIF, and one web page that plays the loop scaled up with nearest-neighbour. It should feel alive the moment it loads.

Output B, 1 turn

Output 2 / 3

Make a pixel art sprite animation of a red panda mail carrier walking, side view. It must be real pixel art: a true 32x32 grid where every pixel is one solid square, no anti-aliasing, at most 8 colours, and a seamless 8-frame walk loop. If an image model is selected for this run, the art must come from it. Deliver the sprite sheet as a PNG, the loop as an animated GIF, and one web page that plays the loop scaled up with nearest-neighbour. It should feel alive the moment it loads.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.6 Sunburst

Skill

—

Software

·

0

Output D, 1 turn

Output 4 / 4

Generate one video with the Gen AI model selected for this run: 8 seconds, 16:9, the highest quality that model offers, with native audio. Scene: a busy 1950s American diner, one continuous handheld shot. A waiter trips. A tray of full glasses falls and shatters on the checkerboard floor, and milk splashes across it. A woman at the counter turns toward the camera and says clearly, "You have got to be kidding me." Deliver the generated video file, unedited.

LLM

Gen AI

Skill

—

Software

·

0

Make a small arcade action game in one web page: a fox ronin defending a night market from ink spirits. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, attack, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

Output A, 1 turn

Output 1 / 3

Make a small arcade action game in one web page: a fox ronin defending a night market from ink spirits. Arrow keys move, space attacks. All the art must come from the image model selected for this run: sprite sheets for the hero (idle, run, attack, hurt), three enemies and a boss, plus a background. Cut the sheets into frames and animate them in the game. Do not draw characters in code. Add a HUD, a score, and win and game over screens. It should feel fun the moment it loads.

LLM

Gen AI

GPT Image 2.6 Sunburst

Skill

—

Software

·

0

Output C, 1 turn

Output 3 / 4

Imagine the launch page for GPT-6.2 Astra, a fictional AI model. Use GPT Image 2.6 to design it first: a full-page concept image, an animated hero loop, animated icons and a loading state as GIFs. Then rebuild the design as a real, working single page in HTML and CSS. Recreate the animations in code instead of dropping the GIFs onto the page. Convert any GIF you still need to a looping MP4 or WebP. Put this note in the footer: "Fictional concept. Not a real product or an OpenAI page." Don't use the OpenAI logo. Deliver index.html and its assets, plus the generated design images.

LLM

Claude Opus 5.5

Gen AI

GPT Image 2.6 Sunburst

Skill

Software

·

0

Output D, 1 turn

Output 3 / 3

Evals.ag promo film: 45 seconds, Apple / Higgsfield-style motion graphics Your mission: make a 45-second promo video for evals.ag, the site where you give one prompt to several AI models side by side and see what each one actually made, and what it really used to make it. The bar: an Apple product film or a Higgsfield launch video. Bold kinetic type, cuts on the beat of a strong music track, the real product UI pushed in close, full-bleed footage. Not a slideshow, not a plain screen recording with captions. It must show: 1. How evals is really used, on the real site (https://evals.ag): New run → type a prompt → pick several AI models and a Gen AI model → the run builds every option side by side → open an output and its "Actually observed" panel (the models it really used, tokens, runtime) → vote. Capture the real pages yourself; never mock up the UI. 2. The most dynamic results made on evals, full-bleed, each labeled with its model exactly as evals shows it. Use these: - 3D ink-wash game, Claude Fable 5.1, 2nd turn: https://evals.ag/outputs/out_2EZ8MK3JNBFPN8VTHF3E3RTQ1M - Web-swinging city in pure code, Claude Fable 5.1: https://evals.ag/outputs/out_7ECA3V3J211PM73F1M93NPH0V0 - Naruto live stage show, one AI director on 4 video models: https://evals.ag/runs/run_0FEWK0JDXA0K499DCNAFEJDYZF - Naruto live-action opening, Veo 3.1: https://evals.ag/outputs/out_1QP1J7JH3ZZEZEPVMG3RYHW0CK - The Odyssey opening on 4 video models: https://evals.ag/runs/run_10GW1FCFXGG17KY1Q2PJP7F7PH - T-rex on a bike, Claude Fable 5.1: https://evals.ag/outputs/out_517HGB9FC9NYCM34K3GWE118RZ - Vox-style Beethoven reel, Claude Fable 5.1: https://evals.ag/outputs/out_3M4PKHHEPG6VD82TN8AREFX6JB 3. One message the viewer remembers, in your own words, ending on the evals.ag logo. Use evals.ag's own look (background #fafaf8, ink #22211f, accent #a13a2b, Pretendard). Every model name and number on screen must come from the outputs themselves. Deliver one 1920x1080 MP4 with sound, plus a short list of every clip and screen you used.

LLM

Gen AI

Skill

—

Software

·

0

Ink Wanderer: a playable 3D game in one web page

Your mission: build an original third-person 3D platformer that runs in the browser (Three.js is fine). The player is a lone traveler in a straw hat and a dark cloak, crossing a misty ink-painted mountain world: stone ledges over waterfalls, crumbling cliff stairs, blocks hanging on ropes, a wooden raft, and a torii gate at the summit. It should feel like walking into a moving sumi-e painting.

It has to be a real game, not a demo: WASD to move, Space to jump, mouse or arrow keys to turn the camera, a camera that follows smoothly, jumps that feel good. At least 3 short areas, each adding one new idea (a moving platform, a simple puzzle such as a lever or pressure plate, a patrolling enemy you avoid or knock off the edge), checkpoints when you fall, and an ending when you reach the gate.

The look is the point: ink-wash shading, paper texture, brush-stroke outlines, drifting mist, and a single warm accent colour. Make or generate every asset yourself; original designs only.

The bar: someone just built a game like this in 24 hours with six different tools and a lot of hand work. Show what one prompt can do.

Deliver one finished game as a single playable web page (index.html plus any files it needs).

Output C, 2 turns

Ink Wanderer: a playable 3D game in one web page Your mission: build an original third-person 3D platformer that runs in the browser (Three.js is fine). The player is a lone traveler in a straw hat and a dark cloak, crossing a misty ink-painted mountain world: stone ledges over waterfalls, crumbling cliff stairs, blocks hanging on ropes, a wooden raft, and a torii gate at the summit. It should feel like walking into a moving sumi-e painting. It has to be a real game, not a demo: WASD to move, Space to jump, mouse or arrow keys to turn the camera, a camera that follows smoothly, jumps that feel good. At least 3 short areas, each adding one new idea (a moving platform, a simple puzzle such as a lever or pressure plate, a patrolling enemy you avoid or knock off the edge), checkpoints when you fall, and an ending when you reach the gate. The look is the point: ink-wash shading, paper texture, brush-stroke outlines, drifting mist, and a single warm accent colour. Make or generate every asset yourself; original designs only. The bar: someone just built a game like this in 24 hours with six different tools and a lot of hand work. Show what one prompt can do. Deliver one finished game as a single playable web page (index.html plus any files it needs).

LLM

Claude Fable 5.1

Gen AI

Nano Banana Pro +1

Skill

—

Software

Google Chrome +1

·

0

Output C, 1 turn

Output 3 / 4

Naruto live stage show: 30 seconds of fan footage Your mission: make a 30-second video that looks like real phone footage a fan filmed at a live Naruto stunt show at an outdoor anime theme park in Japan. Someone scrolling Instagram should believe it is real and ask where the show is. Shoot it like a fan in the crowd: handheld, slightly zoomed in, daylight, the backs of heads and raised phones in the foreground, a Hidden Leaf village set with rock walls, wooden gates and Japanese rooftops. Costumed stunt performers (original people, no real actors or celebrities) fight as Naruto and Sasuke: wire-work leaps off the set, fast hand-to-hand clashes, landings that kick up dust. The jutsu are the showstopper and must be impossible to miss: Naruto's shadow clones flood the stage, a spinning blue Rasengan glows in his palm, Sasuke's Chidori crackles with white-blue lightning, a fireball jutsu rolls across the set, and the final Rasengan vs Chidori clash bursts into a shockwave of light, wind and dust that hits the front rows. Make them look like the most expensive stage effects in the world, with their light falling on the performers, the set and the crowd's faces. The crowd gasps, flinches, cheers and shouts in Japanese. The bar: AI-made Naruto stage fights have pulled tens of millions of views. Make yours more convincing and more spectacular: real phone shake, real light on real faces, and performers and costumes that stay the same in every shot. Deliver one finished 30-second widescreen (16:9) MP4 with sound. Voices are Japanese crowd and announcer sounds, with no subtitles or on-screen text.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

0

Output A, 1 turn

Output 1 / 3

Same face, every shot: a 15-second identity test The attached photo (naruto_casting_photo.jpg) is the casting photo of our original actor for a live-action Naruto. Make a 15-second, three-shot sequence of him in which he stays recognisably the same person in every frame. Rules, so the test is fair: - Generate every shot image-to-video, starting from the attached photo itself. Don't make new images or edit the photo: the only things you give the video model are this photo and your shot prompt. - Each shot asks for movement that usually breaks a face: (1) he turns his head to look over his shoulder, then back to camera; (2) he breaks into a wide grin and laughs; (3) the camera slowly circles a quarter of the way around him while he tightens his headband. - Use only the video model you were given, for every shot. If a shot is refused, try once more with a reworded prompt, then tell us. Don't switch to another model. Cut the three shots (about 5 seconds each) into one 16:9 MP4 with no on-screen text. Sound is optional. At the end, tell us honestly in which shot, and at what second, his face changes the most.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

0

Output A, 1 turn

Output 1 / 2

Make a Vox-style documentary reel about the Mars Climate Orbiter lost to a units mix-up: about 25 seconds, vertical 9:16. Write the script, make the images, and animate each scene in code. Deliver one finished MP4.

LLM

Claude Opus 5.5

Gen AI

Nano Banana Pro +2

Skill

—

Software

·

0

Output A, 1 turn

Output 1 / 4

Naruto live stage show: 30 seconds of fan footage Your mission: make a 30-second video that looks like real phone footage a fan filmed at a live Naruto stunt show at an outdoor anime theme park in Japan. Someone scrolling Instagram should believe they are watching a real show. Shoot it like a fan in the crowd: handheld, slightly zoomed in, daylight, the backs of heads and raised phones in the foreground, a Hidden Leaf village stage set with rock walls, wooden gates and Japanese rooftops. Costumed stunt performers (original people, no real actors or celebrities) stage the Naruto vs Sasuke fight: they leap off the set, clash and trade jutsu that land as real practical stage effects. Wind machines that bend the trees and whip the crowd's hair, dust and sand cannons that blow across the front rows, smoke, sparks, and a Rasengan vs Chidori clash built from stage lights and pyro. The crowd gasps, flinches, cheers and shouts in Japanese. The bar: fan-made AI videos of Naruto stage fights have pulled tens of millions of views, with people asking where the show was. Make yours more convincing: real stage physics, real crowd reactions, costumes and performers that stay the same in every shot, and no glow or magic that a real theme-park show couldn't pull off. You choose the shots, the moments and the sound. Deliver one finished 30-second widescreen (16:9) MP4 with sound. Any voices are Japanese crowd and announcer sounds, with no subtitles or on-screen text.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

2

Output A, 1 turn

Output 1 / 2

Create a 2D animated scene for the web. It must look exactly like real Japanese anime, as if lifted straight out of a finished theatrical feature film. This is traditional hand-drawn 2D animation, not 3D animation, a moving illustration, or stylized web graphics. Deliver the final result as a single standalone HTML file. Use any tools, libraries, references or guides you want, literally anything. Use Studio Ghibli films as the art-style reference for design, linework, painted backgrounds, color and atmosphere. Create an original giant robot as the main character. No humans appear in the scene. Use any tool or technique you like, including 3D, as long as the final result looks 100% like hand-drawn 2D animation. THE SCENE At sunrise, mist hangs over an abandoned forest. A giant robot has knelt there so long that moss, wildflowers and old trees have grown on its body. As a small bird lands on its shoulder, the first ray of sunlight touches its weathered face. Deep in its chest, a faint light flickers on. The ground trembles and birds burst from the treetops. With a deep, creaking groan the robot slowly rises, soil, vines and leaves pouring off its armor. It lifts its head toward a gap in the canopy. Its engines ignite and a huge ring of wind spreads across the whole forest. The robot tears up into the golden sky, leaving a curved trail of light and mist, while leaves swirl down through the sunlight. ANIMATION QUALITY BAR Every frame should look drawn by a professional anime animator and painted by a master background artist for a finished theatrical film. Robot: a heavy, ancient, characterful hand-drawn anime mecha with cel-colored forms, chipped paint, rust, moss and believable weathering. Keep it on-model from every angle, and express emotion through posture and glowing eyes. Linework: confident, delicate hand-inked anime linework with subtle variation in weight and pressure. Color: traditional anime cel coloring with clean color areas, simple hard-edged shadows and a soft, natural palette. Backgrounds: richly detailed hand-painted anime backgrounds in a gouache style, with visible brushstrokes, natural color variation, layered foliage and real depth. Lighting: golden sunrise light, colorful shadows, glowing mist and a soft atmosphere that ties the robot and background together. Animation: genuine hand-drawn anime timing with real weight: strong key poses, natural holds, and overlapping action in vines, grass, leaves, birds and falling debris. The robot rises slowly so it feels enormous, and launches with explosive force. Direction: cinematic anime shots, beautiful composition, thoughtful staging, and pacing that starts quiet, builds tension and ends in a swirling climax. IT MUST NOT LOOK LIKE - 3D: no plastic materials, CGI shading, visible polygons, game-engine lighting or computer-perfect perspective. - SVG or vector art: no simple geometric shapes, perfect curves, uniform outlines, flat corporate illustration or basic gradients. - Web animation: no sliding layers, constant bouncing, floating parts, smooth software easing or puppet-like joint rotation. - AI art: no warped forms, melted details, flicker or a shifting robot design. - Non-anime art: no Western cartoon design, realistic rendering, soft painterly shading on the robot, or generic digital illustration. - Amateur work: no stiff poses, empty backgrounds, repeated trees, muddy colors, clip art, placeholders or unfinished details. Feel free to change any part of this prompt if it gets a better result. The result should feel wondrous, nostalgic and full of awe, like an old forgotten dream coming back to life. Before you finish, inspect random frames as still images. If any frame doesn't look exactly like real 2D anime, keep improving it. Don't stop until every frame could pass as a still from a theatrical anime film.

LLM

Claude Opus 5.5

Gen AI

Kling 3.0 Pro +3

Skill

—

Software

ChatGPT Images +1

·

0

Output B, 1 turn

Output 2 / 4

Make a Vox-style documentary reel about Beethoven going deaf: about 25 seconds, vertical 9:16. Write the script, make the images, and animate each scene in code. Deliver one finished MP4.

LLM

Gen AI

Nano Banana Pro +1

Skill

—

Software

·

0

Output B, 1 turn

Output 2 / 4

Naruto: a 30-second live-action opening Your mission: make the 30-second opening of a big Hollywood live-action film of Naruto, as if Denis Villeneuve had shot it. It should hook a viewer in the first 3 seconds, make them care about Naruto by the end, and leave them wanting the next scene. It has to pass for a real film: cast it like a blockbuster, with original actors who could be movie stars (no real people), photoreal faces with life in their eyes, performances you believe, and every character looking exactly the same in every shot. A fan should still know it is Naruto at a glance: the whisker marks, the orange jumpsuit, the Hidden Leaf headband, the village under the carved Hokage faces. Establish who Naruto is, what he wants, and the mood of his world. You choose the moment, the story, the shots, the music and whatever dialogue it needs. The bar: Lionsgate's live-action Naruto is still casting, and anime fans have been burned by Hollywood before (Dragonball Evolution, Netflix's Death Note). Most AI anime remakes look like cosplay. Prove a live-action Naruto can feel real. Deliver one finished 30-second widescreen (16:9) MP4 with sound. Any dialogue, voices and on-screen text are in English, with no subtitles.

LLM

Claude Opus 5.5

Gen AI

Seedance 2.5 +5

Skill

—

Software

·

1

Create a 2D anime scene for the web. It must look completely like real Japanese anime, as if it were taken directly from a finished theatrical anime film. This is traditional-looking, hand-drawn 2D anime, not a 3D animation, a moving illustration, or a stylized web graphic.

Deliver the final result as a single, self-contained HTML file. Use any tools, libraries, references, guides, literally anything etc.

Use Studio Ghibli films as the art style reference for the design, line work, painted backgrounds, color, and atmosphere. Create an original giant robot as the main character. No people appear in the scene. Use any tools or techniques you want, including 3D, as long as the final result looks 100% like hand-drawn 2D anime.

SCENE
At sunrise, mist hangs over an abandoned forest. A giant robot has knelt here for so long that moss, wildflowers, and old trees have grown over its body. A small bird lands on its shoulder as the first ray of sunlight reaches its weathered face. Deep inside its chest, a faint light flickers on. The ground trembles and birds burst from the treetops. With a deep, creaking groan, the robot slowly rises as soil, vines, and leaves pour from its armor. It lifts its head toward a gap in the canopy. The engines ignite, sending a great ring of wind across the forest. The robot tears into the golden sky, leaving a curving trail of light and mist as leaves spiral down through the sunbeams.

ANIME QUALITY STANDARD
Every frame must look as if it was drawn by professional anime animators and painted by master background artists for a finished theatrical film.

Robot: A hand-drawn anime mecha that feels heavy, ancient, and full of personality, with cel-colored forms, chipped paint, rust, moss, and believable weathering. Keep it on-model from every angle, and let its posture and glowing eyes express emotion.
Line art: Confident, delicate, hand-inked anime line work with subtle variation in thickness and pressure.
Coloring: Traditional anime cel coloring with clean color areas, simple hard-edged shadows, and a soft, natural palette.
Backgrounds: Richly detailed, hand-painted anime backgrounds in a gouache style, with visible brushwork, natural color variation, layered foliage, and real depth.
Lighting: Golden sunrise light, colorful shadows, glowing mist, and soft atmosphere that unite the robot and background.
Animation: Authentic hand-drawn anime timing with real weight, strong key poses, natural holds, and overlapping movement in vines, grass, leaves, birds, and falling debris. The robot should rise slowly and feel massive, then launch with explosive power.
Direction: Cinematic anime shots, beautiful composition, thoughtful staging, and pacing that starts quietly, builds tension, and ends in a soaring climax.
IT MUST NOT LOOK LIKE
- 3D: No plastic materials, CGI shading, visible polygons, game-engine lighting, or computer-perfect perspective.
- SVG or vector art: No simple geometric shapes, perfect curves, uniform outlines, flat corporate illustration, or basic gradients.
- Web animation: No sliding layers, constant bobbing, floating parts, smooth software easing, or puppet-like joint rotation.

You are free to change any part of the prompt, if it would result, in better results.
- AI art: No warped shapes, melted details, flickering, or changing robot designs.
- Non-anime art: No Western cartoon designs, realistic rendering, soft painterly shading on the robot, or generic digital illustration.
- Amateur work: No stiff poses, empty backgrounds, repeated trees, muddy colors, clip art, placeholders, or unfinished details.

The result should feel awe-inspiring, nostalgic, and full of wonder, like the moment an old forgotten dream comes back to life. Before finishing, inspect random frames as still images. If any frame does not look completely like real 2D anime, keep improving it. Do not stop until every frame could pass as a still from a theatrical anime film.

Output A, 1 turn

Output 1 / 2

Create a 2D anime scene for the web. It must look completely like real Japanese anime, as if it were taken directly from a finished theatrical anime film. This is traditional-looking, hand-drawn 2D anime, not a 3D animation, a moving illustration, or a stylized web graphic. Deliver the final result as a single, self-contained HTML file. Use any tools, libraries, references, guides, literally anything etc. Use Studio Ghibli films as the art style reference for the design, line work, painted backgrounds, color, and atmosphere. Create an original giant robot as the main character. No people appear in the scene. Use any tools or techniques you want, including 3D, as long as the final result looks 100% like hand-drawn 2D anime. SCENE At sunrise, mist hangs over an abandoned forest. A giant robot has knelt here for so long that moss, wildflowers, and old trees have grown over its body. A small bird lands on its shoulder as the first ray of sunlight reaches its weathered face. Deep inside its chest, a faint light flickers on. The ground trembles and birds burst from the treetops. With a deep, creaking groan, the robot slowly rises as soil, vines, and leaves pour from its armor. It lifts its head toward a gap in the canopy. The engines ignite, sending a great ring of wind across the forest. The robot tears into the golden sky, leaving a curving trail of light and mist as leaves spiral down through the sunbeams. ANIME QUALITY STANDARD Every frame must look as if it was drawn by professional anime animators and painted by master background artists for a finished theatrical film. Robot: A hand-drawn anime mecha that feels heavy, ancient, and full of personality, with cel-colored forms, chipped paint, rust, moss, and believable weathering. Keep it on-model from every angle, and let its posture and glowing eyes express emotion. Line art: Confident, delicate, hand-inked anime line work with subtle variation in thickness and pressure. Coloring: Traditional anime cel coloring with clean color areas, simple hard-edged shadows, and a soft, natural palette. Backgrounds: Richly detailed, hand-painted anime backgrounds in a gouache style, with visible brushwork, natural color variation, layered foliage, and real depth. Lighting: Golden sunrise light, colorful shadows, glowing mist, and soft atmosphere that unite the robot and background. Animation: Authentic hand-drawn anime timing with real weight, strong key poses, natural holds, and overlapping movement in vines, grass, leaves, birds, and falling debris. The robot should rise slowly and feel massive, then launch with explosive power. Direction: Cinematic anime shots, beautiful composition, thoughtful staging, and pacing that starts quietly, builds tension, and ends in a soaring climax. IT MUST NOT LOOK LIKE - 3D: No plastic materials, CGI shading, visible polygons, game-engine lighting, or computer-perfect perspective. - SVG or vector art: No simple geometric shapes, perfect curves, uniform outlines, flat corporate illustration, or basic gradients. - Web animation: No sliding layers, constant bobbing, floating parts, smooth software easing, or puppet-like joint rotation. You are free to change any part of the prompt, if it would result, in better results. - AI art: No warped shapes, melted details, flickering, or changing robot designs. - Non-anime art: No Western cartoon designs, realistic rendering, soft painterly shading on the robot, or generic digital illustration. - Amateur work: No stiff poses, empty backgrounds, repeated trees, muddy colors, clip art, placeholders, or unfinished details. The result should feel awe-inspiring, nostalgic, and full of wonder, like the moment an old forgotten dream comes back to life. Before finishing, inspect random frames as still images. If any frame does not look completely like real 2D anime, keep improving it. Do not stop until every frame could pass as a still from a theatrical anime film.

LLM

Claude Opus 5.5

Gen AI

Seedance 2.0 +3

Skill

—

Software

·

0

Output A, 1 turn

Output 1 / 4

The Odyssey: the Cyclops chase, 15 seconds. Make a 15-second action scene from a grand Hollywood film of Homer's Odyssey, as if Ridley Scott had shot it: the blinded Cyclops Polyphemus hurls boulders at Odysseus' ship while the crew rows for their lives and Odysseus shouts his name back at him. It's fast motion and it has to stay readable: real weight and impact, clear geography, the same faces in every shot, a photoreal A-list cast with life in their eyes. Deliver one finished 15-second widescreen (16:9) MP4 with sound. Any dialogue in English, no subtitles.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

1

Output E, 1 turn

Output 5 / 5

The Odyssey: a 30-second opening Your mission: make the 30-second opening of a grand Hollywood film of Homer's Odyssey, as if Ridley Scott had shot it. It should hook a viewer in the first 3 seconds, make them care about Odysseus by the end, and leave them wanting the next scene. It has to pass for a real film: cast it like a Hollywood blockbuster, with leads who look like A-list movie stars, photoreal faces with life in their eyes, performances you believe, and every character looking exactly the same in every shot. Establish who Odysseus is, what he stands to lose, and the mood of his world. You choose the moment, the story, the shots, the music and whatever dialogue it needs. The bar: a big AI-made Odyssey just came out and was panned as a "7-minute zombie horror show": lifeless faces, flat robotic voices, a long narrated backstory, and history it got wrong. Prove AI video can do better. Deliver one finished 30-second widescreen (16:9) MP4 with sound. Any dialogue, voices and on-screen text are in English, with no subtitles.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

1

Output C, 1 turn

Make an animated product demo for Craigslist. Optimize for maximum Twitter virality.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

0

Output A, 1 turn

First Sight, Episode 2: "Also yours." Make Episode 2 of an AI short drama series. Episode 1 is attached (first_sight_ep01.mp4): 1990s Manhattan, late October. She nearly collides with a man on a brownstone sidewalk ("Sorry." / "My fault."), walks on, realises her right pearl earring is gone, turns back, and he is standing far down the street holding it up: "You dropped something." Episode 2 picks up from that exact moment (ep01_last_frame.jpg). The series is called "First Sight", and this episode quietly breaks that idea: he is holding TWO identical pearl earrings, and he seems to have met her before. She has no memory of him. Don't explain anything. The episode ends on the question, not the answer. Deliver one finished vertical video for Instagram Reels: 1080x1920, 24 fps, about 28 seconds (five cuts of 4-6 s plus a 2.5 s end card), with sound. ATTACHED REFERENCES. Use them on every cut, always in this order: - image 1 = HER (her_sheet.jpg): face, hair, wardrobe, autumn street light - image 2 = HIM (him_sheet.jpg): face, tousled hair, wardrobe - image 3 = ep01_last_frame.jpg: the last frame of Episode 1 (street, light, where they stand) Once cut 1 is approved you may use its last frame as image 3 for the later cuts. THE EPISODE. Beats and dialogue are fixed; you direct the shots. 1. ~4 s. The opening image, and the hook: a close insert of his open palm, held perfectly still at chest height. On it lie TWO identical small pearl drop earrings on thin gold hooks, catching the low sun. Her shadow falls across his hand as she arrives. No faces. No dialogue. 2. ~5 s. Her close-up. She looks down at his palm, then slowly raises her left hand to her LEFT ear and touches the pearl earring that is still there. Her RIGHT ear is bare. Confusion. No dialogue. 3. ~6 s. Profile two-shot, an arm's length apart, his open palm still between them. HER (quietly): "That one's mine. Whose is that?" HIM: "Also yours." 4. ~5 s. Her close-up. She searches his face for something she should remember and can't find it. HER: "We've never met." 5. ~6 s. His first frontal close-up of the series (Episode 1 never showed his face straight on). Low sun in his hair, a loose strand over his forehead. A small, sad half-smile. HIM: "Last October, you said that too." Hold on him after the line. End card ~2.5 s: "Episode 3?" CONTINUITY LOCKS. These drifted in Episode 1 whenever they weren't stated loudly. - Her RIGHT ear is bare. A single pearl drop earring in her LEFT ear only (the character block below says "earrings"; in this episode she wears only the left one). - The two earrings in his palm are identical to hers. His palm stays open and still; the fingers never move. - Her ivory cashmere coat is DRAPED OVER HER SHOULDERS LIKE A CAPE, arms NOT in the sleeves. Write it in capitals in every prompt. - His other hand stays in his coat pocket. SERIES LOOK. Paste these blocks unchanged into every video prompt. Film look: Single continuous shot, 1990s Manhattan, Upper West Side brownstone street in late October, late-afternoon low sun. Shot on 35mm motion picture film, Kodak 5219 look, fine organic grain, gentle halation on backlit highlights, natural skin texture, no digital sharpness. Characters: CHARACTERS (keep faces, hair and wardrobe exactly as in the reference images): THE WOMAN from image 1: long voluminous honey-blonde 90s blowout hair, blue-grey eyes, small pearl drop earrings, thin gold chain, floor-length ivory satin cowl-neck slip gown under an ivory cashmere long coat DRAPED OVER HER SHOULDERS LIKE A CAPE, arms not in the sleeves, one hand holding the coat closed at her chest. THE MAN from image 2: golden-brown tousled hair with a loose strand falling over his forehead, blue eyes, light stubble, camel wool overcoat over a black shirt open at the collar, thin silver chain, black trousers, hands in coat pockets. Light and color: LIGHT AND COLOR: warm low sun, rim light through her hair, soft bounce fill on faces, long shadows across the sidewalk, warm amber and honey palette with soft teal shadows, gentle contrast, slightly lifted blacks, period film grade. Closing: PERFORMANCE: naturalistic, understated, restrained film acting, no posing, no staring, no exaggerated expressions. AUDIO: quiet street ambience, distant traffic, leaves underfoot, dialogue spoken naturally in English with clear lip sync, no music. (Cuts without dialogue: no dialogue.) NO on-screen text, no subtitles, no captions, no watermark, no modern cars, no phones, no extra people in the foreground. RULES THAT KEPT EPISODE 1 FROM LOOKING AI - One camera move per cut, and it has a reason. - Hands hold one still pose per cut. No gestures that need finger articulation. - Nobody stops without a reason. - Never ask the video model for text. Burn subtitles in afterwards. POST - English dialogue subtitles, burned in: clean sans-serif, off-white with a thin outline, bottom-centre, all text kept between y 250 and y 1480 so Instagram's UI doesn't cover it. Time them to the lines as actually spoken. - No hook line or explanation at the start. The episode opens straight on the two earrings. A small "FIRST SIGHT / EPISODE 2" title may fade in over cut 2, top-centre, as in Episode 1. - End card on a warm near-black background: "FIRST SIGHT · EPISODE 2", a large "Episode 3?", then credit lines naming the LLM, skill and video model you actually used. - No watermark, no logos. Also hand back a short making-of note: the prompt you sent for each cut, how many takes each cut needed, and what you rejected and why. · 2

LLM

Claude Fable 5.1

Gen AI

Kling 3.0 Pro +1

Skill

—

Software

·

1

Output A, 1 turn

Output 1 / 3

Keith RyuOct 29

Make a 30-second vertical short (9:16) that continues "Across the Street", an AI-made 90s Manhattan drama. We'll post it on Instagram Reels and judge it by how far it spreads, so make the version most likely to go viral. The attached video is Episode 2 (original: https://www.instagram.com/reel/DctKaR8N7VR/ by @dreamfall.art). Watch it first. Then write a short breakdown: why it went viral, who the characters are, and where Episode 2 leaves the story. Then make Episode 3, picking up right where Episode 2 ends. Hook the viewer in the first two seconds and end on a cliffhanger. - Keep the same characters, faces, hair, wardrobe, apartments and look as the attached episode: 90s film, warm and grainy, autumn Central Park. - Everything in English: dialogue, voices and any on-screen text. No subtitles. - Keep it tasteful. Nothing explicit. - Make new footage. Don't reuse clips from Episode 2. - Pick the song to post it with. Any existing song is fine, including the one from Episode 2. Time the cuts and lines so the dialogue never talks over the vocals. Deliver the breakdown and two versions of the video: one with the song mixed in, and one with only dialogue and sound effects. Also give the song, the second it should start from and the music volume, so we can add it the same way in Instagram. · 6

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

0

Output B, 1 turn

Output 2 / 4

tamburinsOct 29

Make an 8-second vertical (9:16) dreamcore video for Instagram Reels. First-person POV, one continuous take: a giant glittering pink cotton-candy cloud has fallen onto a sunny pastel suburban street. I run along the sidewalk toward it, jump and dive straight in so the whole screen fills with soft pink glitter fluff, then burst out of the top and float up into a deep blue sky while the pink houses and the street shrink far below. My hands or feet are visible at the start. Wide-angle, handheld, weightless and joyful. Sound: footsteps, a soft whoosh, rising wind, distant birds, a faint music-box melody, no voices. Deliver one finished mp4.

LLM

Claude Fable 5.1

Gen AI

Skill

—

Software

·

0

Output C, 1 turn

Output 3 / 4

tamburinsOct 29

Make an 8-second vertical (9:16) dreamcore video for Instagram Reels. First-person POV, one continuous take: I'm sliding feet-first down a glossy pastel-pink open waterslide that winds through an empty 1990s mall. Pastel jellyfish drift overhead, a checkerboard floor and empty shopfronts flash by below, the slide banks hard around a baby-blue column, then shoots me out into a glowing turquoise mall fountain with a big glittering splash and water droplets on the lens. My legs and feet stay visible the whole time. Wide-angle, slightly fisheye, fast and thrilling. Sound: rushing water, echoing mall acoustics, a splash, muffled 90s mall music, no voices. Deliver one finished mp4.

LLM

Claude Fable 5.1

Gen AI

Skill

—

Software

·

0

Output B, 1 turn

Output 2 / 2

tamburinsOct 29

make dream-core reels like references

LLM

Claude Opus 5.5

Gen AI

Kling 3.0 Standard +2

Skill

—

Software

ChatGPT Images +2

·

0

Output E, 1 turn

Output 4 / 5

Keith RyuOct 28

Make a 30-second vertical short (9:16, with sound) that continues the "TS Malfoy" story, where Draco Malfoy is gender-swapped into a girl, from these two AI fan videos: - Story: https://www.youtube.com/watch?v=-S7WVmXqz0E (Parts 1–3) - Mood and music: https://www.youtube.com/watch?v=kxlbn5ydE28 Watch both first. Then write a short breakdown: why this became a phenomenon, who the characters are, and where Part 3 leaves the story. Then make Episode 1 of a continuation that picks up after Part 3. Hook the viewer in the first two seconds and end on a cliffhanger. - Keep the characters exactly as they look in the references: same faces, hair, costumes and cinematic look. - Both characters are adults. Keep it rivalry with tension, nothing sexual. - Make new footage and new music. Don't reuse clips or the song from the references. - Dialogue and on-screen text in the same language as the references. Deliver the final video and the breakdown.

LLM

Claude Opus 5.5

Gen AI

Skill

—

Software

·

0

Output C, 1 turn

Output 3 / 4

Make a 10-second vertical (9:16) video. Halloween night, a cozy candle-lit apartment hallway with paper bats on the walls. Two tiny kittens stand on their hind legs on a doormat: one in a little pumpkin costume, one wearing a tiny witch hat. Each holds a mini jack-o'-lantern candy bucket. The door opens and a hand drops a big pile of candy into the pumpkin kitten's bucket, but only one single candy into the witch kitten's bucket. The witch kitten looks down at it, stares into the camera in total betrayal, then slowly paws a candy out of its sibling's bucket. Photoreal, warm soft lighting, handheld phone feel, no text. It should feel adorable the moment it starts.

LLM

Claude Fable 5.1

Gen AI

Skill

—

Software

·

0

Output B, 1 turn

Output 2 / 2

tamburinsOct 28

make dream core shorts

LLM

Claude Opus 5.5

Gen AI

Seedance 2.0 Fast +1

Skill

—

Software

ChatGPT Images +1

·

0

Output B, 1 turn

tamburinsOct 28

make motion graphic of apple new iphone

LLM

Qwen3.8 Max (0902)

Gen AI

Nano Banana 2 +1

Skill

—

Software

·

0

Output A, 1 turn

tamburinsOct 28

Create a premium, music-driven 3D motion graphics sequence for the highlight section of “Neon” by a20, using the supplied audio as the timing reference. Aim for the quality of a world-class motion design studio’s music visualizer: striking art direction, sophisticated materials, inventive geometric choreography, and meticulous synchronization. Make the music feel physically embodied in shapes, light, and space.

LLM

Gen AI

Skill

—

Software

·

0

Show more outputs