Test log
Leonardo.Ai test data
Raw results for the Leonardo.Ai review. The verdict is in the review.
- Tested
- 5 November 2026
- Tester
- Miles Foster
- Account
- Free demo account, signed up by RankDude, nothing paid to the brand
- Tests
- 10
- Screenshots
- 6
- Method
- Published methodology
- Environment
- Google Chrome on desktop
Results
| ID | Test | Key measurement | Score |
|---|---|---|---|
| T01 | Style Range | Styles run: 0 of 10 | n/a |
| T02 | Character & Subject Consistency | Images run: 0 of 12 | n/a |
| T03 | Editing, Inpainting & Image-to-Image | Edits run: 0 of 12 | n/a |
| T04 | Prompt Adherence | Elements correct: 65 of 72 | 5/5 |
| T05 | Photorealism & Anatomy | Images that pass: 3 of 5 | 3/5 |
| T06 | Text in Images | Words correct: 34 of 34 | 4/5 |
| T07 | Resolution & Upscaling | Native size: 1,024 px (1.05 MP) | 4/5 |
| T08 | Speed per Image | Median per image: 10.3 s | 4/5 |
| T09 | Usage Rights, Watermarks & Privacy | Points met: 3 of 8 checked | 2/5 |
| T10 | Pricing & Value | Free images I got in one day: 12, plus one upscale | 5/5 |
| Overall (in the review) | 4.3/5 | ||
Style Range
n/aNot Tested- Method
- Ten style prompts for one lighthouse.
- Styles run
- 0 of 10
- Remark
- Not graded: no style prompt was run.
Raw notes
Ten style prompts for one lighthouse.
Not run: ten images cost 80 tokens and the balance was 36 after the other tests.
- Run on the next review of this tool.
Character & Subject Consistency
n/aNot Tested- Method
- A described woman and a described thermos in six scenes each.
- Images run
- 0 of 12
- Remark
- Not graded: no consistency prompt was run.
Raw notes
A described woman and a described thermos in six scenes each.
Not run: twelve images cost 96 tokens, and the reference feature (Add Image) is marked as a paid feature.
- Run on a plan with the reference feature.
Editing, Inpainting & Image-to-Image
n/aNot Tested- Method
- Twelve fixed edits.
- Edits run
- 0 of 12
- Remark
- Not graded: no edit was run.
Raw notes
Twelve fixed edits.
Not run: a written edit shows a price of 40 tokens and the balance was 36. Adding a reference image or a mask needs a paid plan.
- Run on a plan with tokens to spare.
Prompt Adherence
5/5Tested- Method
- 12 of the 40 fixed prompts (3 portraits, 2 product shots, 3 set-wording images, 4 busy scenes), typed in by script, first image kept, every written element counted at full size.
- Elements correct
- 65 of 72
- Busy scenes
- 20 of 27
- Remark
- Scale 5: 90 percent or more of elements correct. The sample was 12 prompts and 72 elements, not 40 prompts and about 240.
Raw notes
12 of the 40 fixed prompts (3 portraits, 2 product shots, 3 set-wording images, 4 busy scenes), typed in by script, first image kept, every written element counted at full size.
65 of 72 elements were right (90.3 percent). Portraits, products and set-wording images: 45 of 45. The four busy scenes: 20 of 27.
Beach scene 7 of 7. Kitchen scene 5 of 8: four apples instead of three, apples on both sides of the jug, the cat on the windowsill instead of the chair. Living room 4 of 6: three cushions, not four, one white, not two. Desk 4 of 6: three books, not two, and the plant's pot is hidden.
Added content that changed the scene: invented chalkboard text on the night market poster.
- 12 of 40 prompts. The two groups not run (styles and character scenes) are the ones the free plan could not pay for.
Photorealism & Anatomy
3/5Tested- Method
- Five portrait and product images (P1, P2, P3, F1, F4), one run each, inspected at 100 percent for faults that give a generated image away.
- Images that pass
- 3 of 5
- Hand and face faults
- 0
- Remark
- Scale 3: 13 to 17 of 24 pass. The sample is 5 images, so 3 of 5 (60 percent) stands for 14 to 15 of 24.
Raw notes
Five portrait and product images (P1, P2, P3, F1, F4), one run each, inspected at 100 percent for faults that give a generated image away.
Passed (3): the man on the bench, the laughing woman and the girl with the balloon, with no hand or face faults.
Failed (2): the watch dial prints "MAONICY" and "IASEMON" instead of a brand, and the cork on the olive oil bottle is covered in marks like fake letters.
Hand faults 0. Face faults 0. Faults in text-like marks on objects 2.
- Five images in one run, not 24 images in two runs.
Text in Images
4/5Tested- Method
- Three of the six set-wording prompts: the coffee shop sign and board (11 words), the night market poster (13 words) and the handwritten fridge note (10 words), 34 words in all.
- Words correct
- 34 of 34
- Images fully correct
- 2 of 3
- Remark
- Scale 4: four images fully correct. Two of three stands for four of six, with every word right.
Raw notes
Three of the six set-wording prompts: the coffee shop sign and board (11 words), the night market poster (13 words) and the handwritten fridge note (10 words), 34 words in all.
Every one of the 34 words was spelled right, in the place asked for.
Two of three images were fully correct. The poster also carried invented chalkboard text ("Handmade Local Unique You'll" and "GOOD FOOD GOOD VIBES"), so it counts as a fault.
Auto Image used GPT Image 2 for all three; Lucid Origin made the other nine images.
- Three of six prompts; the street sign, book cover and jar label were not run.
Resolution & Upscaling
4/5Tested- Method
- Every saved image was measured, the aspect ratios were read in the app, and one image was upscaled at the largest setting the free plan offers.
- Native size
- 1,024 px (1.05 MP)
- Upscaled size
- 2,048 px
- Remark
- Scale 4: 1 megapixel or more native, an upscale with minor invented texture. The scale's top mark needs 4,096 pixels, which the free plan did not offer.
Raw notes
Every saved image was measured, the aspect ratios were read in the app, and one image was upscaled at the largest setting the free plan offers.
All 12 images came out at 1,024 by 1,024 pixels (1.05 megapixels) as JPEG. The app showed 1,344 by 768 for 16:9 and 832 by 1,248 for 2:3.
The 2x Pro Precise upscale cost 18 tokens and gave 2,048 by 2,048 pixels (4.1 megapixels). Sand and sandcastle came out sharper and cleaner; the child's face gained dark marks that were not in the original.
The app then said "Max size reached" and that free accounts can upscale images up to 5 megapixels.
- One upscale, not six. The free plan stopped at 2x.
Speed per Image
4/5Tested- Method
- Each of the 12 generations was timed from pressing Generate to the full-size image on screen.
- Median per image
- 10.3 s
- Slowest image
- 26.5 s
- Remark
- Scale 4: a median of 10 to 19.9 seconds.
Raw notes
Each of the 12 generations was timed from pressing Generate to the full-size image on screen.
Median 10.3 seconds, average 12.3, fastest 8.2, slowest 26.5.
The three set-wording images (GPT Image 2) took 17.4, 26.5 and 17.4 seconds; the nine Lucid Origin images took 8.2 to 11.3.
No generation failed. There was no queue on the free plan at the time.
- 12 generations, not 40.
Usage Rights, Watermarks & Privacy
2/5Tested- Method
- The terms of service (read 5 November 2026) and the downloaded files were checked against the ten-point list.
- Points met
- 3 of 8 checked
- Remark
- Scale 2: 4 or 5 of 10. 3 of 8 checked is 37 percent, so about 4 of 10.
Raw notes
The terms of service (read 5 November 2026) and the downloaded files were checked against the ten-point list. Eight points could be checked on the free plan.
Met (3): the terms say who owns the output; a paid plan owns what it makes (clause 8.3); the free download had no visible watermark and was the same file as the stored image.
Not met (5): free outputs are owned by Leonardo (clause 8.7) so no commercial use on free; all free images are public; free public content may be used to train models (clause 8.6), and only paid private content is protected (clause 8.5); no stated source for training data; only the 3 GPT Image 2 files and the upscale carried a content credentials label, the 9 Lucid Origin files did not.
Not tested (2): a watermark on paid downloads, and whether real people are refused when asked for.
- Terms read by me, not a lawyer. No paid plan was bought.
Pricing & Value
5/5Tested- Method
- The free tier was used until the app stopped me, and the plans page was read in both billing views.
- Free images I got in one day
- 12, plus one upscale
- Cost per image on Essential
- about $0.011
- Remark
- Scale 5: under $0.02 an image with at least 10 free images a day. The paid per-image price assumes the 8-token cost carries over.
Raw notes
The free tier was used until the app stopped me, and the plans page was read in both billing views.
Free: 150 tokens a day and 8 tokens an image. I got 12 images and one 2x upscale from one day's tokens before the app refused another image with 36 tokens left.
The cheapest plan that grants ownership is Essential at $12 a month ($10 a month paid yearly) for 8,500 tokens. At 8 tokens an image that is about $0.011 an image, if a paid image costs the same.
Costs on top: 18 tokens for a 2x upscale and 40 for a written edit. Private mode, reference images and models other than Auto need a paid plan.
- No plan bought; paid token costs per image were not read.
Coverage
Tested
- 12 of the 40 fixed prompts, first image kept: 3 portraits, 2 product shots, 3 set-wording images, 4 busy scenes
- Time and token cost of every image
- One 2x Pro Precise upscale, opened and compared with the original
- The product's own Download button, opened and compared with the stored file
- Terms of service and the plans page read in full
Not tested or limited
- The free plan's 150 daily tokens paid for 12 images and an upscale: 12 of the 40 prompts were run
- Style range, consistency and editing were not graded: the app said "You don't have enough tokens" at a balance of 36, an edit costs 40 and adding a reference image is a paid feature
- Photorealism used one run of 5 images, not two runs of 12 prompts
- No plan was bought: paid ownership, private generations, the paid file set and cancellation were not tested
- Prompts naming real people or brands were not run, so the enforcement point of the rights checklist was not tested
Still to test
- Style range, consistency and editing on a paid plan
- Photorealism with 24 images, and the other 28 prompts
- The other models Auto Image can choose
- Ownership, private mode and cancellation after buying Essential