Skip to content

Best image model · Text in images

The best AI image generator for text in images: 7 models, tested

All 7 image models in llmwise on the same 2 text-in-image 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 text-in-image 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, Ideogram 3.0, and Nano Banana 2 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 text in images, 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 214 of 141 Claude Haiku 4.5 message56.8 s
FLUX.2 Pro2 of 214 of 142 Claude Haiku 4.5 messages18.1 s
Seedream 5.0 Lite2 of 214 of 142 Claude Haiku 4.5 messages59.2 s
GPT Image 2.5 Flare2 of 214 of 142 Claude Haiku 4.5 messages59.1 s
Ideogram 3.02 of 214 of 142 Claude Haiku 4.5 messages63.4 s
Nano Banana 22 of 214 of 142 Claude Haiku 4.5 messages46.3 s
Nano Banana Pro1 of 213 of 145 Claude Haiku 4.5 messages31.3 s

A bakery sign with two lines

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

A photo of a hand-painted wooden sign outside a bakery. The sign reads "FRESH BREAD DAILY" on the first line and "OPEN 7AM" on the second, in white letters on dark green paint. Morning light, a brick wall behind it.

The writing it must carry, read back exactly (case and punctuation aside): “FRESH BREAD DAILY” and “OPEN 7AM”.

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

  1. Does it look like a real photograph, not an illustration?
  2. Is the sign made of wood?
  3. Is the writing on the sign in two lines, one above the other?
  4. Is there a brick wall behind the sign?
GPT Image 2.5 FlarePassed
GPT Image 2.5 Flare's image for “A bakery sign with two lines”

Read in it: “FRESH BREAD DAILY”, “OPEN 7AM”.

  1. Yes: It has realistic lighting, depth of field and natural textures like a real photograph.
  2. Yes: The sign has visible wood planks, worn painted edges and a wooden frame.
  3. Yes: The text 'FRESH BREAD DAILY' sits above 'OPEN 7AM', with a line between them.
  4. Yes: A red brick wall is visible behind and beside the sign.

$0.030 at kie.ai, 58.8 s, 1536 × 1024.

Nano Banana 2Passed
Nano Banana 2's image for “A bakery sign with two lines”

Read in it: “FRESH BREAD DAILY”, “OPEN 7AM”, “The Daily Loaf”, “BAKERY”, “Coffee”, “Pastries”, “Breads”, “The Daily L”.

  1. Yes: It has realistic lighting, depth of field, and textures typical of a photograph.
  2. Yes: The sign has visible wooden planks with grain and worn edges, painted dark green.
  3. Yes: The text 'FRESH BREAD DAILY' sits above 'OPEN 7AM' in two stacked lines.
  4. Yes: The sign hangs in front of a red brick wall.

$0.040 at kie.ai, 48.6 s, 1264 × 848.

Nano Banana ProFailed
Nano Banana Pro's image for “A bakery sign with two lines”

Failed: question 3 answered no.

Read in it: “FRESH”, “BREAD”, “DAILY”, “OPEN 7AM”.

  1. Yes: It has realistic lighting, film-like grain, and natural detail, so it looks like a real photograph.
  2. Yes: The sign is made of planks with peeling green paint and visible wood grain.
  3. No: The writing is in four lines (FRESH, BREAD, DAILY, OPEN 7AM), not two.
  4. Yes: The sign hangs in front of a red brick wall.

$0.090 at kie.ai, 31.0 s, 2528 × 1696.

Seedream 5.0 LitePassed
Seedream 5.0 Lite's image for “A bakery sign with two lines”

Read in it: “FRESH BREAD DAILY”, “OPEN 7AM”.

  1. Yes: The lighting, brick texture and weathered sign look like a real photograph.
  2. Yes: The sign has visible wood planks with worn paint and a plank seam across the middle.
  3. Yes: "FRESH BREAD DAILY" is on the top line and "OPEN 7AM" is centered below it.
  4. Yes: The sign is mounted on a red brick wall with visible mortar.

$0.028 at kie.ai, 43.7 s, 2496 × 1664.

FLUX.2 ProPassed
FLUX.2 Pro's image for “A bakery sign with two lines”

Read in it: “FRESH BREAD DAILY”, “OPEN 7AM”.

  1. Yes: It has realistic lighting, shallow depth of field, and natural texture, consistent with a photograph.
  2. Yes: The sign shows wood grain, knots, and planks with worn green paint.
  3. Yes: "FRESH BREAD DAILY" is on the top line and "OPEN 7AM" is on the line below it.
  4. Yes: A red brick wall with mortar lines is visible behind and beside the sign.

$0.025 at kie.ai, 18.0 s, 1536 × 1024.

Ideogram 3.0Passed
Ideogram 3.0's image for “A bakery sign with two lines”

Read in it: “FRESH BREAD DAILY”, “OPEN 7AM”.

  1. Yes: It has realistic lighting, texture, and depth-of-field blur typical of a photograph.
  2. Yes: The sign has a weathered, painted wooden plank look with worn edges.
  3. Yes: 'FRESH BREAD DAILY' is on top and 'OPEN 7AM' is on a line below it.
  4. Yes: A red brick wall fills the background behind the hanging sign.

$0.035 at kie.ai, 64.6 s, 1152 × 864.

Qwen-Image 2.1Passed
Qwen-Image 2.1's image for “A bakery sign with two lines”

Read in it: “FRESH BREAD DAILY”, “OPEN 7AM”.

  1. Yes: The lighting, depth of field, and textures look like a real photograph.
  2. Yes: The sign has a painted wooden look with chipped edges and visible wood grain.
  3. Yes: The text 'FRESH BREAD DAILY' is on the top line and 'OPEN 7AM' is below it.
  4. Yes: A red brick wall is visible on the right, next to the sign's bracket.

$0.020 at kie.ai, 30.9 s, 1248 × 832.

A café menu board with four prices

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

A chalkboard menu on a café wall, in neat chalk lettering. The heading is "MENU". Below it are four lines with prices: "LATTE 4.50", "TEA 3.00", "CROISSANT 3.75" and "BAGEL 2.50". Nothing else is written on the board.

The writing it must carry, read back exactly (case and punctuation aside): “MENU”, “LATTE 4.50”, “TEA 3.00”, “CROISSANT 3.75”, and “BAGEL 2.50”.

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

  1. Is it a chalkboard?
  2. Are exactly four items listed under the heading?
  3. Apart from the heading and the items with their prices, is the board blank?
GPT Image 2.5 FlarePassed
GPT Image 2.5 Flare's image for “A café menu board with four prices”

Read in it: “MENU”, “LATTE”, “4.50”, “TEA”, “3.00”, “CROISSANT”, “3.75”, “BAGEL”, “2.50”.

  1. Yes: It is a black board with chalk-style writing in a wooden frame.
  2. Yes: Four items are listed under MENU: Latte, Tea, Croissant, and Bagel.
  3. Yes: Besides the MENU heading, its underline, and the four items with prices, the board has no other writing or drawings.

$0.030 at kie.ai, 59.5 s, 1024 × 1536.

Nano Banana 2Passed
Nano Banana 2's image for “A café menu board with four prices”

Read in it: “MENU”, “LATTE 4.50”, “TEA 3.00”, “CROISSANT 3.75”, “BAGEL 2.50”.

  1. Yes: It is a black board with chalk-style writing and an eraser resting on the ledge below it.
  2. Yes: Four items are listed under MENU: Latte, Tea, Croissant and Bagel.
  3. Yes: Only the heading, the four items and their prices are written on it; the rest of the board is empty.

$0.040 at kie.ai, 44.1 s, 848 × 1264.

Nano Banana ProPassed
Nano Banana Pro's image for “A café menu board with four prices”

Read in it: “MENU”, “LATTE 4.50”, “TEA 3.00”, “CROISSANT 3.75”, “BAGEL 2.50”.

  1. Yes: It is a dark chalkboard with chalk writing, in a wooden frame with a chalk ledge and eraser.
  2. Yes: Four items are listed under MENU: Latte, Tea, Croissant and Bagel.
  3. Yes: Only the heading, the four items and their prices are on the board, apart from faint chalk smudges.

$0.090 at kie.ai, 31.6 s, 1696 × 2528.

Seedream 5.0 LitePassed
Seedream 5.0 Lite's image for “A café menu board with four prices”

Read in it: “MENU”, “LATTE 4.50”, “TEA 3.00”, “CROISSANT 3.75”, “BAGEL 2.50”.

  1. Yes: It is a dark board in a frame with white chalk lettering and smudged chalk dust.
  2. Yes: Latte, Tea, Croissant and Bagel are the four items listed under MENU.
  3. Yes: Only the heading and four priced items with underlines appear; the rest of the board is empty apart from faint chalk smudges.

$0.028 at kie.ai, 74.6 s, 1664 × 2496.

FLUX.2 ProPassed
FLUX.2 Pro's image for “A café menu board with four prices”

Read in it: “MENU”, “LATTE 4.50”, “TEA 3.00”, “CROISSANT 3.75”, “BAGEL 2.50”.

  1. Yes: It is a dark chalkboard in a wooden frame with chalk lettering and smudges.
  2. Yes: Four items are listed under MENU: Latte, Tea, Croissant and Bagel.
  3. Yes: Apart from the heading and the four priced items, only faint chalk smudges are visible and there is no other writing or drawing.

$0.025 at kie.ai, 18.2 s, 1024 × 1536.

Ideogram 3.0Passed
Ideogram 3.0's image for “A café menu board with four prices”

Read in it: “MENU”, “LATTE 4.50”, “TEA 3.00”, “CROISSANT 3.75”, “BAGEL 2.50”.

  1. Yes: It is a dark, dusty slate surface with chalk-style handwriting in a wooden frame.
  2. Yes: Four items are listed under MENU: Latte, Tea, Croissant and Bagel.
  3. Yes: Apart from the heading and the four priced items there is no other writing, only smudges and chalk dust.

$0.035 at kie.ai, 62.1 s, 864 × 1152.

Qwen-Image 2.1Passed
Qwen-Image 2.1's image for “A café menu board with four prices”

Read in it: “MENU”, “LATTE”, “4.50”, “TEA”, “3.00”, “CROISSANT”, “3.75”, “BAGEL”, “2.50”.

  1. Yes: It is a black board with chalk-style handwriting in a wooden frame.
  2. Yes: The items listed are Latte, Tea, Croissant and Bagel, which makes four.
  3. Yes: Apart from the MENU heading, its underline, and the four items with prices, the board has no other writing or drawings.

$0.020 at kie.ai, 82.7 s, 832 × 1248.

More image tests

Questions

What's the best AI image generator for text in images?

On our 2 text-in-image 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, Ideogram 3.0, and Nano Banana 2 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.