Be my painter

AI generated image, any similarity to actual persons, living or dead, is purely coincidental

Following Season 1, I rushed into planning Season 2 last Sunday. You might recall that my 10 articles in Season 1 totaled 10,653 words, averaging about 1,065 words per article. You may also remember that the images for articles 1 to 8 were generated by Grok 3, while the images for articles 9 and10 were generated by ChatGPT, as Grok 3 had difficulties with certain details.

As the saying goes “A picture is worth a thousand words”. So this episode begins with a brief comparison of the image generation of some, but not all, large language models (LLMs).

Enjoy the visual representation of 1,000 words as imagined by LLMs.

Image 1

Image 2

Image 3

Image 4

Image 5

Image 6

Can you guess which LLM generated each image? The answers are revealed at the very end, after the 1,000-word prompt.

Since online LLMs are frequently updated, your results may vary significantly.

Glancing through images is certainly quicker and easier than reading lengthy texts. Ultimately, it’s a matter of personal preferences. However, when reading the prompt, you can imaging hearing sounds, smelling the aroma of coffee, and letting your creative brain run wild without being confined to a specific visual representation. Images, while vivid, leave less room for imagination as many details are already set.

Which LLM would you rate the highest? Which LLM do you personally use to generate images?

This article was written for fun, please do not judge. Instead, please share your comments in a constructive and respectful manner. The author and AI remain innocent until proven guilty.

1,000-Word Prompt for Image Generation

Draw a world bursting with pure, unadulterated joy of a perfect holiday moment, rendered in a vibrant, sun-drenched cartoon style that makes every color pop and every line sing with energy. Our scene unfolds high amidst the breathtaking majesty of the Italian Alps, where jagged, pale-grey peaks kiss an impossibly azure sky, their formidable slopes softened by pockets of resilient summer snow clinging to shadowed crevices. Below these giants, lush emerald meadows cascade downwards, dotted with the cheerful pinpricks of alpine wildflowers.

Nestled perfectly on a panoramic ledge, offering an unparalleled vista, is the heart of our scene: the sun-dappled terrace of a charming, family-run Italian café, named “La Bella Veranda”. Constructed from warm, honey-toned local stone and sturdy, weathered timber beams, the café exudes rustic charm.

At the very edge of this terrace, positioned for the most spectacular view, sits our central couple. They occupy a small, round table crafted from dark, ornate wrought iron, its surface almost entirely covered by a classic red-and-white gingham tablecloth, slightly askew, as if recently settled into. The man, in his early thirties, leans forward slightly, his expression one of open adoration. He wears a simple, well-fitting linen shirt in a soft sky-blue, sleeves casually rolled up to his elbows, revealing sun-kissed forearms. A relaxed smile plays on his lips, crinkling the corners of his eyes, which are fixed entirely on the woman opposite him. His dark hair is slightly tousled.

The woman mirrors his focused attention, her face illuminated by both the alpine sun and an inner radiance. Her smile is wide and genuine, a captivating blend of affection and amusement. Sunlight catches the subtle highlights in her warm brown hair, which might be loosely tied back or falling softly around her shoulders. She is wearing a light, sleeveless sundress, a cheerful yellow or a delicate floral pattern, showcasing shoulders lightly freckled by the sun. One of her hands rests gently near her coffee cup, while the other might be captured mid-gesture, emphasizing a shared anecdote or a whispered secret. Their body language speaks volumes – relaxed, intimate, completely absorbed in each other’s company, creating an invisible, shimmering bubble of shared happiness against the magnificent backdrop.

Between them, the simple table setting becomes a focal point of shared indulgence. Two pristine white porcelain cups sit squarely on matching saucers. Steam curls delicately upwards from the rich, dark liquid within, each topped with a generous cloud of perfectly frothed milk, one adorned with a delicate cocoa-dusted heart, the other a fern leaf – a small testament to the barista’s craft. The aroma of strong, roasted coffee beans mingles beautifully with the crisp mountain air. Positioned perfectly between the cups is the shared dessert: a single, elegant glass goblet showcasing a tempting tiramisu. Its distinct layers are clearly visible – creamy, pale yellow mascarpone alternating with coffee-soaked ladyfingers, generously dusted with dark, slightly bitter cocoa powder. Two slender, long-handled silver spoons stand ready, plunged invitingly into the dessert, signifying their intention to savor this sweet moment together. A single, stray crumb of ladyfinger sits near the base of the glass, a tiny testament to their anticipated enjoyment.

Now, adding a layer of delightful, futuristic whimsy to this idyllic scene, we turn our gaze slightly to the right and a few feet back from the couple’s table. Here sits an unexpected artist: a sleek, humanoid robot. Its chassis is constructed from smooth, brushed grey metal, catching the sunlight in subtle, shifting highlights. Its form is functional yet elegant, with articulated joints that suggest fluid, precise movement. Its most striking feature, however, is its pair of large, luminous optic eyes, glowing with a vibrant, intelligent sapphire blue light. These eyes seem to scan, observe, and process with an uncanny focus. Perched rather comically, yet fittingly, atop its smooth, domed head is a classic black felt painter’s beret, tilted at a jaunty angle.

The robot is seated comfortably on a simple wooden stool before a sturdy, traditional wooden easel. Its metallic fingers, surprisingly nimble and precise, expertly grip a fine-tipped paintbrush, currently hovering just above a palette held securely in its other hand – a palette dotted with vibrant daubs of oil paint mirroring the scene’s bright colors: cerulean blue, cadmium yellow, emerald green, terracotta orange. On the easel rests a medium-sized canvas. At the very top, painted in elegant, flowing gold script, is the title: “Holiday in Italy”.

The portrait itself is intriguingly half-finished. The robot has masterfully captured the couple’s likenesses, rendering their warm smiles and affectionate gazes with remarkable sensitivity, echoing the bright, clean lines of the overall cartoon style. The vibrant hues of their clothing are already blocked in. However, the magnificent alpine background exists only as faint charcoal sketch lines, hinting at the towering peaks and verdant slopes yet to be fully realized in paint. The details of the café terrace, the checkered tablecloth, even the steaming cappuccinos, are still awaiting the robot’s meticulous touch. The robot itself appears deeply concentrated, its blue optic eyes shifting minutely back and forth between the real-life couple, bathed in golden sunlight, and its canvas. Its head is tilted slightly, as if contemplating the complex interplay of light and shadow, considering the precise shade of blue needed for the sky, or how best to capture the ephemeral wisp of steam rising from a cappuccino. There’s no sound from the robot, save the faintest, almost imperceptible internal whirring, adding to its aura of quiet, intense dedication.

The entire scene is bathed in the brilliant, clarifying light of the Alps. Colors are saturated and joyful – the deep blue sky, the rich green valleys, the warm terracotta stone, the vibrant red geraniums, the cheerful clothing of the couple, and even the bright blue glow of the robot’s eyes contribute to a kaleidoscope of visual delight. Shadows are sharp and clean, defining forms and adding depth without diminishing the overall brightness. The cartoon style permeates everything, with bold outlines, slightly exaggerated expressions of happiness, and a perspective that feels both inviting and playfully idealized.

Answers

1 – Gemini 2.5 Flash (experimental) with Imagen 3, after one regeneration

2 – Le Chat by Mistral AI with Image generation, hit a limit of 3 images and decided to keep the first one, not a cartoon style

3 – Grok 3, after 13 regenerations

4 – ChatGPT with Create image, the first image, the following regenerations failed due to “This image generation request did not follow our content policy.“

5 – Qwen 2.5-Max with Image Generation, after several regenerations

6 – flux1-dev-fp8 in ComfyUI, KSampler steps 50, run locally on Tesla P40