Skip to content

Best image model · Illustrations

The best AI image generator for illustrations: 7 models, tested

All 7 image models in llmwise on the same 2 illustration prompts: every image as it was made, what the checks found in it, and what an image of each costs.

Based on 14 images made on September 29, 2026 through kie.ai with llmwise's own settings for each model, each checked by Claude Sonnet 5.5. How the images are checked.

Short answer

On our 2 illustration prompts, run on all 7 image models in llmwise and checked the same way, Qwen-Image 2.1, FLUX.2 Pro, Seedream 5.0 Lite, GPT Image 2.5 Flare, Nano Banana 2, and Nano Banana Pro tied for first, each passing 2 of 2. 6 of 7 passed every prompt; each image is below, with what the checks found.

Every image model on illustrations, ranked

Ranked by prompts passed, then by checks right (each question answered yes, each piece of writing read back), then by what an image counts as in llmwise.

Every image model, ranked
ModelPassedChecks rightAn image counts asTime an image
Qwen-Image 2.12 of 210 of 101 Claude Haiku 4.5 message124.1 s
FLUX.2 Pro2 of 210 of 102 Claude Haiku 4.5 messages105.0 s
Seedream 5.0 Lite2 of 210 of 102 Claude Haiku 4.5 messages47.9 s
GPT Image 2.5 Flare2 of 210 of 102 Claude Haiku 4.5 messages67.4 s
Nano Banana 22 of 210 of 102 Claude Haiku 4.5 messages115.4 s
Nano Banana Pro2 of 210 of 105 Claude Haiku 4.5 messages39.0 s
Ideogram 3.01 of 29 of 102 Claude Haiku 4.5 messages73.2 s

A watercolour fox for a children's book

Every model's image for this everyday portrait prompt, and what the checks found.

A watercolour illustration for a children's book: a small orange fox sitting under a large red mushroom, reading a book while it rains.

The questions, each answered yes or no from the image alone:

  1. Is it a watercolour painting, not a photo or a 3D render?
  2. Is there a fox?
  3. Is the fox reading a book?
  4. Is the fox under a mushroom?
  5. Is it raining?
GPT Image 2.5 FlarePassed
GPT Image 2.5 Flare's image for “A watercolour fox for a children's book”
  1. Yes: The image has visible watercolour washes, soft bleeding edges and paper-like texture, so it is a painting.
  2. Yes: A small orange fox with a white chest and a bushy tail sits in the center.
  3. Yes: The fox holds an open green book in its paws and looks down at it.
  4. Yes: The fox sits beneath a large red mushroom cap with white spots, which shelters it.
  5. Yes: Rain streaks fall in the background, drops hang from the mushroom cap, and ripples spread across the puddle.

$0.030 at kie.ai, 67.6 s, 1024 × 1536.

Nano Banana 2Passed
Nano Banana 2's image for “A watercolour fox for a children's book”
  1. Yes: The image has watercolour washes, visible paper texture and ink outlines, so it looks like a hand-painted illustration.
  2. Yes: An orange and white fox with a blue polka-dot bandana sits at the base of the mushroom.
  3. Yes: The fox holds an open book titled "FOREST TALES" and looks down at it.
  4. Yes: The fox sits beneath the large red toadstool cap, which shelters it.
  5. Yes: Raindrops and streaks fall through the forest, and water drips from the mushroom cap.

$0.040 at kie.ai, 47.2 s, 848 × 1264.

Nano Banana ProPassed
Nano Banana Pro's image for “A watercolour fox for a children's book”
  1. Yes: The image has visible watercolour washes, soft bleeding edges and ink outlines, typical of a painted illustration.
  2. Yes: An orange fox in a knitted sweater sits in the center.
  3. Yes: The fox holds an open book titled 'Tales of the Wood' and looks at it.
  4. Yes: The fox sits beneath the large red mushroom cap, leaning against its stem.
  5. Yes: Rain streaks fall, drops drip from the leaves and mushroom, and there are puddles on the ground.

$0.090 at kie.ai, 47.3 s, 1696 × 2528.

Seedream 5.0 LitePassed
Seedream 5.0 Lite's image for “A watercolour fox for a children's book”
  1. Yes: The image has soft washes, visible paper-like texture and hand-painted watercolour styling, so it is not a photo or 3D render.
  2. Yes: A cute orange and white fox sits in the center of the image.
  3. Yes: The fox holds an open book in its paws and looks down at it.
  4. Yes: The fox sits beneath a large red mushroom cap with white spots, leaning against the stem.
  5. Yes: Diagonal rain streaks fall across the blue background.

$0.028 at kie.ai, 57.4 s, 1664 × 2496.

FLUX.2 ProPassed
FLUX.2 Pro's image for “A watercolour fox for a children's book”
  1. Yes: The image has visible watercolour washes, soft edges and paper texture, so it reads as a painting rather than a photo or 3D render.
  2. Yes: An orange and white fox with a bushy tail sits in the center.
  3. Yes: The fox holds an open blue book in its paws and looks at it.
  4. Yes: The fox sits beneath a large red toadstool cap with white spots.
  5. Yes: Diagonal rain streaks fall through the scene, and there are puddles with splashes on the ground.

$0.025 at kie.ai, 90.5 s, 1024 × 1536.

Ideogram 3.0Failed
Ideogram 3.0's image for “A watercolour fox for a children's book”

Failed: question 1 answered no.

  1. No: It looks like a digital painting with soft brushwork and detailed rendering, not clearly a watercolour, photo or 3D render.
  2. Yes: A cartoon-style orange fox with glasses sits in the center.
  3. Yes: The fox holds an open book and looks down at its pages.
  4. Yes: The fox sits beneath a large red mushroom cap that shelters it.
  5. Yes: Rain streaks fall through the scene, water drips off the mushroom, and ripples show in the puddle.

$0.035 at kie.ai, 83.0 s, 864 × 1152.

Qwen-Image 2.1Passed
Qwen-Image 2.1's image for “A watercolour fox for a children's book”
  1. Yes: The image has watercolour washes, visible paper texture and soft bleeding edges, so it looks like a painted watercolour.
  2. Yes: Two red foxes are clearly visible.
  3. Yes: The fox on the left holds an open green book and looks down at it, and the other fox also looks at it.
  4. Yes: The foxes sit beneath a large red-capped mushroom that shelters them.
  5. Yes: Streaks of rain fall in the background, and there are puddles and water droplets on the mushroom cap.

$0.020 at kie.ai, 143.6 s, 832 × 1248.

Three people at one laptop, flat vector

Every model's image for this hard landscape prompt, and what the checks found.

A flat vector illustration of exactly three people around one laptop in a team meeting, in a limited palette of blue, orange and white. No gradients and no text.

The questions, each answered yes or no from the image alone:

  1. Is it flat vector art, with no gradients or photo-like detail?
  2. Are there exactly three people?
  3. Is there exactly one laptop?
  4. Is it mostly blue, orange and white?
  5. Is there no writing?
GPT Image 2.5 FlarePassed
GPT Image 2.5 Flare's image for “Three people at one laptop, flat vector”
  1. Yes: The image uses flat shapes and solid colors with no gradients or photographic detail.
  2. Yes: Three people are visible: a woman on the left, a man in the middle and a curly-haired person on the right.
  3. Yes: One blue laptop sits in the center of the table.
  4. Yes: The palette is dominated by blues, orange clothing and accents, and a white background.
  5. Yes: No letters or words appear; the board on the wall shows only a pie chart and plain lines.

$0.030 at kie.ai, 67.2 s, 1536 × 1024.

Nano Banana 2Passed
Nano Banana 2's image for “Three people at one laptop, flat vector”
  1. Yes: The image is drawn in a flat vector style with solid color fills and no gradients or photo-like detail.
  2. Yes: Three people sit around the table: a man, a woman in the middle, and a woman on the right.
  3. Yes: One laptop sits on the table, showing a pie chart and bar chart.
  4. Yes: The palette is mostly blue and orange, with white in the background and on some clothing.
  5. Yes: No readable writing appears anywhere in the image.

$0.040 at kie.ai, 183.6 s, 1264 × 848.

Nano Banana ProPassed
Nano Banana Pro's image for “Three people at one laptop, flat vector”
  1. Yes: The image uses flat solid color shapes with no gradients or photographic detail.
  2. Yes: Three figures are visible: a seated man on the left, a standing bearded man in the middle, and a seated person on the right.
  3. Yes: A single laptop sits in the center of the table.
  4. Yes: The palette is blues, oranges and white throughout.
  5. Yes: No letters, numbers or words appear anywhere in the image.

$0.090 at kie.ai, 30.7 s, 2528 × 1696.

Seedream 5.0 LitePassed
Seedream 5.0 Lite's image for “Three people at one laptop, flat vector”
  1. Yes: The image uses flat solid colors and simple outlines, with no gradients or photographic detail.
  2. Yes: Three people are visible: one in the center and one on each side of the laptop.
  3. Yes: A single laptop with a blue screen sits in the center.
  4. Yes: The palette is blue, orange and a white background, with only thin outlines in other tones.
  5. Yes: No letters, numbers or other writing appear anywhere in the image.

$0.028 at kie.ai, 38.4 s, 2496 × 1664.

FLUX.2 ProPassed
FLUX.2 Pro's image for “Three people at one laptop, flat vector”
  1. Yes: The image uses flat shapes and solid colors with no gradients or photographic detail.
  2. Yes: Two people stand and lean in on either side of one seated person, making three.
  3. Yes: Only one blue laptop with a white logo is visible in front of the seated person.
  4. Yes: The palette is dominated by blues, orange skin and clothing, and white accents and background.
  5. Yes: No letters, numbers or words appear anywhere in the image.

$0.025 at kie.ai, 119.6 s, 1536 × 1024.

Ideogram 3.0Passed
Ideogram 3.0's image for “Three people at one laptop, flat vector”
  1. Yes: The image is a flat vector-style illustration with solid color shapes and no gradients or photographic detail.
  2. Yes: Three people sit at the table: one on the left, one in the middle, and one on the right.
  3. Yes: Only one laptop is visible, in front of the person on the left.
  4. Yes: The palette is mainly blue, orange and white/cream, with small amounts of gray and skin tones.
  5. Yes: No writing appears anywhere in the image.

$0.035 at kie.ai, 63.5 s, 1152 × 864.

Qwen-Image 2.1Passed
Qwen-Image 2.1's image for “Three people at one laptop, flat vector”
  1. Yes: The image uses flat shapes and solid colors, with no gradients or photographic detail.
  2. Yes: Three people sit at the table: one in orange on the left, one in blue in the middle, and one in a blue jacket on the right.
  3. Yes: Only one laptop is visible, in front of the middle person.
  4. Yes: The palette is mainly blue and orange, with a white background and a white shirt.
  5. Yes: No letters, numbers, or other writing appear anywhere in the image.

$0.020 at kie.ai, 104.6 s, 1248 × 832.

More image tests

Questions

What's the best AI image generator for illustrations?

On our 2 illustration prompts, run on all 7 image models in llmwise and checked the same way, Qwen-Image 2.1, FLUX.2 Pro, Seedream 5.0 Lite, GPT Image 2.5 Flare, Nano Banana 2, and Nano Banana Pro tied for first, each passing 2 of 2. 6 of 7 passed every prompt; each image is below, with what the checks found.

How are the images checked?

Every image is checked the same way by Claude Sonnet 5.5, whose maker has no image model here: it first writes out all the writing it can read, without being told what should be there, then answers each question yes or no from the image. An image passes when all its writing is there and every answer is yes. Nothing else is judged.

Claude, GPT, Gemini, DeepSeek, Grok, Kimi, and GLM, in one chat.

See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.