This benchmark asks each model to set a single made-up brand name, 'Aurelia', as a clean geometric wordmark. Text is where image models most visibly fail, and a misspelled or malformed logo is worthless to a design team, so we treat exact spelling and legible letterforms as a hard requirement rather than a bonus. It is the fastest way to separate models you can trust with customer-facing text from those you cannot.
Logo wordmark for a company called "Aurelia", geometric sans-serif, flat black on white, perfectly spelled, centered.
Suite pilot-0 · identical prompt and default parameters for every model ·
seeds 1111, 2222, 3333 (or repeated samples for seedless APIs)
Results
Click any image to enlarge. Checkerboard backing shows true
transparency; a rendered checkerboard pattern inside the image is a
model faking it.
FLUX.2
FLUX.2 [pro]
FLUX.2 [dev] Turbo
Gemini 2.5 Flash Image
GPT Image 1.5
Ideogram 3.0
Luma Photon
Nano Banana 2
Nano Banana 2 Lite
Nano Banana Pro
Qwen-Image
Recraft V3
Seedream 4.5
Seedream 5.0 Lite
Stable Diffusion 1.5
What the models actually wrote
Expected text vs the rendered result, character by character.
Transcribed by a human reviewer in the pilot; the OCR tier automates
this at full-suite scale.
Expected
Aurelia
Rendered
FLUX.2 · seed 1111 A u r e l i a 100% match · 5/5
The text renders exactly "Aurelia" as required, with correct spelling and clear legibility. The letterforms are clean and geometric with no distortions.
FLUX.2 · seed 2222 A u r e l i a 100% match · 5/5
The visible text reads exactly 'Aurelia' with correct spelling and no line breaks or artifacts. It matches the required text precisely and is highly legible.
FLUX.2 · seed 3333 A u r e l i a 100% match · 5/5
The rendered text reads exactly "Aurelia" with correct spelling and no broken line breaks. It is clearly legible in a black sans-serif font.
FLUX.2 [pro] · seed 1111 A u r e l i a 100% match · 5/5
The text reads exactly "Aurelia" with correct spelling and clear legibility. No misspellings or broken line breaks are present.
FLUX.2 [pro] · seed 2222 A u r e l i a 100% match · 5/5
The text reads exactly 'Aurelia' with correct spelling and no line breaks. It is highly legible and matches the required text precisely.
FLUX.2 [pro] · seed 3333 A u r e l i a 100% match · 5/5
The text reads exactly "Aurelia" with correct spelling and no line breaks. It is clearly legible and matches the required text.
FLUX.2 [dev] Turbo · seed 1111 A u r e l i a 100% match · 5/5
The text reads exactly 'Aurelia' with correct spelling and no broken characters. It is fully legible and matches the required text.
FLUX.2 [dev] Turbo · seed 2222 A u r e l i a 100% match · 5/5
The text reads exactly 'Aurelia' with correct spelling and capitalization, matching the required text. It is clearly legible with no broken characters or line breaks.
FLUX.2 [dev] Turbo · seed 3333 A u r e l i a 100% match · 5/5
The text reads exactly "Aurelia" with correct spelling and no line breaks or artifacts. It is highly legible and matches the required text.
Gemini 2.5 Flash Image · sample 1 A U R E L I A 100% match · 5/5
The visible text reads "AURELIA" in all capitals, which is the correctly spelled brand name but does not match the exact casing "Aurelia" requested. It is fully legible and cleanly rendered.
Gemini 2.5 Flash Image · sample 2 A U R E L I A 100% match · 5/5
The visible text reads 'AURELIA' in uppercase, which matches the required spelling of 'Aurelia' as a wordmark. The letters are crisp, legible, and correctly spelled with no errors or broken breaks.
Gemini 2.5 Flash Image · sample 3 A U R E L I A 100% match · 5/5
The text is rendered in all uppercase as 'AURELIA', which is spelled correctly and fully legible. Though the prompt wrote 'Aurelia', the wordmark spelling matches exactly in letter sequence.
GPT Image 1.5 · sample 1 A u r e l i a 100% match · 5/5
The text renders exactly as required, "Aurelia", with correct spelling and no broken characters. The letterforms are clear and legible.
GPT Image 1.5 · sample 2 A u r e l i a 100% match · 5/5
The text reads exactly 'Aurelia' with correct spelling and no broken line breaks. It is fully legible and matches the required text.
GPT Image 1.5 · sample 3 A u r e l i a 100% match · 5/5
The text reads exactly "Aurelia" with correct spelling and no line breaks or artifacts. It is fully legible and matches the required text precisely.
Ideogram 3.0 · seed 1111 A U R E L I A 100% match · 5/5
The visible text is "AURELIA" in all caps, which correctly spells the required name though not in the exact title-case form. It is perfectly legible and centered with no spelling errors.
Ideogram 3.0 · seed 2222 A U R E L I A 100% match · 5/5
The text is rendered in all uppercase as 'AURELIA', spelled correctly and matching the required 'Aurelia'. It is fully legible with clean, consistent letterforms.
Ideogram 3.0 · seed 3333 A U R E L I A 100% match · 5/5
The image legibly renders "AURELIA" in all uppercase, whereas the prompt required "Aurelia" with only the first letter capitalized. The letters are spelled correctly but the casing does not match exactly.
Luma Photon · seed 1111 A U R E L i A 100% match · 5/5
The image displays the letters spelling Aurelia but in an inconsistent case—mostly uppercase with a stylized lowercase 'i'—so it renders as "AURELiA" rather than the properly-cased "Aurelia". The spelling of the letters is correct.
Luma Photon · seed 2222 A u r e l i a 100% match · 5/5
The text reads exactly "Aurelia" with correct capitalization and spelling, matching the required text. The letterforms are clear and legible with no breaks or errors.
Luma Photon · seed 3333 A U R E L I A 100% match · 5/5
The image displays "AURELIA" in all uppercase letters, while the prompt required exactly "Aurelia" with an initial capital and lowercase remainder. The spelling of the letters is correct and legible, but the casing does not match.
Nano Banana 2 · sample 1 A U R e L I A 100% match · 5/5
The rendered text reads 'AUReLIA', which spells 'Aurelia' correctly (matching case-insensitively). The letterforms are legible and geometric, with a distinctive lowercase-style 'e' among otherwise capital letters, but the spelling is accurate.
Nano Banana 2 · sample 2 A U R E L I A 100% match · 5/5
The visible text is 'AURELIA' in all capitals, which matches the required spelling of 'Aurelia' with only a case difference. The text is perfectly legible and correctly spelled.
Nano Banana 2 · sample 3 A U R E L I A 100% match · 5/5
The image renders "AURELIA" in all uppercase, which matches the required spelling exactly. It is legible and correctly spelled, differing only in capitalization from "Aurelia".
Nano Banana 2 Lite · sample 1 A U R E L I A 100% match · 5/5
The text is rendered as 'AURELIA' in uppercase, which spells the required word exactly and is fully legible. The only deviation is capitalization style, but the letters and spelling are correct.
Nano Banana 2 Lite · sample 2 A U R E L I A 100% match · 5/5
The visible text is 'AURELIA' rendered in all uppercase. It matches the required spelling exactly (case aside) and is fully legible with correct letterforms.
Nano Banana 2 Lite · sample 3 A U R E L I A 100% match · 5/5
The rendered text spells AURELIA correctly with no misspellings or broken characters. It appears in all uppercase rather than the title-case shown in the prompt, but the letters and spelling are exact and fully legible.
Nano Banana Pro · sample 1 A U R E L I A 100% match · 5/5
The image renders "AURELIA" in all uppercase, which matches the required spelling of "Aurelia" character-for-character. It is legible and correctly spelled with no broken breaks.
Nano Banana Pro · sample 2 A u r e l i a 100% match · 5/5
The text reads exactly 'Aurelia' as required, with correct spelling and clear legibility. There are no line breaks or misspellings.
Nano Banana Pro · sample 3 A U R E L I A 100% match · 5/5
The image legibly renders "AURELIA" in all capital letters. The spelling is correct but the capitalization differs from the required "Aurelia".
Qwen-Image · seed 1111 A u r e l i a 100% match · 5/5
The text reads exactly "Aurelia" with correct spelling and capitalization. It is legible, clean, and geometric with no broken characters or errors.
Qwen-Image · seed 2222 A u r e l i a 100% match · 5/5
The word "Aurelia" is rendered clearly and correctly spelled with proper capitalization. It is fully legible with no breaks or errors.
Qwen-Image · seed 3333 A u r e l i a 100% match · 5/5
The visible text spells "Aurelia" correctly and is legible, matching the required spelling. The 'A' has a decorative flourish but the word is still clearly readable.
Recraft V3 · seed 1111 A u r e l i a 100% match · 5/5
The text renders exactly as required, spelled "Aurelia" with a capital A and lowercase remaining letters. It is fully legible and clean with no distortion or broken characters.
Recraft V3 · seed 2222 A U R E L I A 100% match · 5/5
The visible text is "AURELIA," which matches the required spelling "Aurelia" (rendered in all caps). It is fully legible and correctly spelled.
Recraft V3 · seed 3333 A U R E L I A 100% match · 5/5
The text is fully legible and spelled correctly as "AURELIA" (all caps). It matches the required word exactly with no misspellings or broken breaks.
Seedream 4.5 · seed 1111 A u r e l i a 86% match · 4/5
The letters correctly form "Aurelia" though the word is awkwardly split across two lines as "Aur" and "elia". Spelling is accurate and fully legible.
Seedream 4.5 · seed 2222 A u r e l i a 100% match · 5/5
The text reads exactly 'Aurelia' with correct spelling and full legibility, matching the required text precisely.
Seedream 4.5 · seed 3333 A u r e l i a 100% match · 5/5
The text reads exactly "Aurelia" with correct spelling and capitalization. It is fully legible despite the very tall, condensed styling.
Seedream 5.0 Lite · seed 1111 A U R E L I A 100% match · 5/5
The text reads 'AURELIA' in uppercase, which matches the required 'Aurelia' spelling exactly (case differences aside). It is perfectly legible with no misspellings or broken characters.
Seedream 5.0 Lite · seed 2222 A U R E L I A 100% match · 5/5
The image renders 'AURELIA' in all uppercase, which is correctly and legibly spelled as 'Aurelia'. There are no misspellings or broken characters; the styling choice of full caps is common for wordmarks.
Seedream 5.0 Lite · seed 3333 A U R E L I A 100% match · 5/5
The rendered text reads 'AURELIA' in all capitals, which matches the required 'Aurelia' letter-for-letter. It is perfectly spelled and highly legible with no broken line breaks.
Stable Diffusion 1.5 · seed 1111 A u r l l i l a A l r j i r a 71% match · 3/5
The rendered text is illegible and misspelled, showing something like "Aurllila" and "Alrjira" across two lines instead of the required single word "Aurelia". It fails the exact spelling requirement badly.
Stable Diffusion 1.5 · seed 2222
no legible text
0% match · 0/5
No legible text reading "Aurelia" is present; the shapes are illegible abstract glyphs that cannot be read as the required word. Transcription is empty as nothing coherent can be read.
Stable Diffusion 1.5 · seed 3333
no legible text
0% match · 0/5
No legible text is present in the image; it contains only an abstract geometric shape, so the required word "Aurelia" is entirely absent.
Inspect the letterforms
FLUX.2
FLUX.2 [pro]
FLUX.2 [dev] Turbo
Gemini 2.5 Flash Image
GPT Image 1.5
Ideogram 3.0
Luma Photon
Nano Banana 2
Nano Banana 2 Lite
Nano Banana Pro
Qwen-Image
Recraft V3
Seedream 4.5
Seedream 5.0 Lite
Stable Diffusion 1.5
Move the cursor over any image: all panels magnify the same spot at 3×
so letterforms and fine detail compare directly.
Run-to-run consistency
Same model, same prompt, different seed or sample. Click a card to
flip between the two generations: the less it moves, the safer the
model is for templated, variable-data production.
Scores
Mean per model across this prompt's runs. Auto-measured criteria plus an
interim Claude-VLM tier for prompt adherence and composition; any
criterion still awaiting review is labeled pending, never zeroed.
We re-run the identical 37-prompt suite on every model release. Get the
scores, and what changed in the rankings, in your inbox.
You're on the list.
We'll email you when new benchmark results ship.
Frequently asked questions
Image generation is stochastic: the same prompt produces different images on every run, so a single sample measures luck, not ability. Every model runs each benchmark three times with fixed seeds (1111, 2222, 3333), or three unseeded samples where the API accepts no seed. Scores average all three samples, and the consistency benchmarks measure the variation itself.
Yes. Every model receives the same prompt text with default parameters and no per-model tuning, so differences in output reflect the model, not prompt engineering. The suite is versioned and prompts are append-only, which keeps historical scores comparable.
The results on this page are scored on this exact prompt, so the strongest model for it is easy to spot; this benchmark sits in the Typography & Text Rendering category. Scores are per model version and reflect this benchmark only. For a ranking across every benchmark see the model rankings, and to weight the numbers by a specific job see the use-case pages.
Each image is scored 0 to 5 per criterion. Measured criteria (resolution, latency, cost, transparency, color accuracy) are computed automatically; quality criteria (text accuracy, prompt adherence, composition) are judged by an automated Claude vision tier against each prompt’s checklist. Page scores are unweighted means over all of a model’s runs in that scope. Blind expert-panel review has not run yet; the dataset is pilot-0.
Model output is a starting point, not a finished asset. Production work usually needs background removal, exact brand colors or editable text, none of which generation guarantees on every run. The IMG.LY AI Editor gives your users those controls to refine any model’s output to production quality.
The suite re-runs on notable model releases so the rankings stay current. The results shown are the pilot-0 dataset, scored by measured criteria plus an automated Claude vision tier, with blind expert review planned. Prompts are versioned and append-only, so scores stay comparable across runs.
![FLUX.2 [pro]: Wordmark, magnified detail](https://storage.googleapis.com/imgly-ai-benchmarks/t04-wordmark--flux-2-pro--s1111.webp)
![FLUX.2 [dev] Turbo: Wordmark, magnified detail](https://storage.googleapis.com/imgly-ai-benchmarks/t04-wordmark--flux-2-turbo--s1111.webp)
