Test log
PicLumen test data
Raw results for the PicLumen review. The verdict is in the review.
- Tested
- 6 November 2026
- Tester
- Miles Foster
- Account
- Free demo account, signed up by RankDude, nothing paid to the brand
- Tests
- 10
- Screenshots
- 7
- 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 97 | 3/5 |
| T05 | Photorealism & Anatomy | Images that pass: 5 of 8 | 3/5 |
| T06 | Text in Images | Words correct: 12 of 50 | 1/5 |
| T07 | Resolution & Upscaling | Native size: 1,024 px (1.05 MP) | 3/5 |
| T08 | Speed per Image | Median per image: 10.9 s | 4/5 |
| T09 | Usage Rights, Watermarks & Privacy | Points met: 3 of 8 checked | 2/5 |
| T10 | Pricing & Value | Free images a day, first week: 20 | 5/5 |
| Overall (in the review) | 3.4/5 | ||
Style Range
n/aNot Tested- Method
- Ten style prompts for one lighthouse.
- Styles run
- 0 of 10
- Remark
- No grade without a run.
Raw notes
Ten style prompts for one lighthouse.
Not run: ten images cost 10 Lumens and the balance was 3 after the other tests.
- Left out.
Character & Subject Consistency
n/aNot Tested- Method
- A described woman and a described thermos in six scenes each.
- Images run
- 0 of 12
- Remark
- No grade without a run.
Raw notes
A described woman and a described thermos in six scenes each.
Not run: twelve images cost 12 Lumens and the balance was 3 after the other tests.
- Left out.
Editing, Inpainting & Image-to-Image
n/aNot Tested- Method
- Twelve fixed edits.
- Edits run
- 0 of 12
- Remark
- No grade without a run.
Raw notes
Twelve fixed edits.
Not run: Edit, Describe, Inpaint, Colorize and Outpaint carry the paid mark on the free plan, and the plans table shows a cross for Basic on Edit, Inpaint and Outpaint.
- Left out.
Prompt Adherence
3/5Tested- Method
- 17 of the 40 fixed prompts (4 portraits, 4 product shots, 5 set-wording images, 4 busy scenes), typed in by script, first image kept, every written element counted at full size.
- Elements correct
- 65 of 97
- Busy scenes
- 10 of 28
- Remark
- Scale 3: 65% to 79% of elements correct. The sample was 17 prompts and 97 elements, not 40 prompts and about 240.
Raw notes
17 of the 40 fixed prompts (4 portraits, 4 product shots, 5 set-wording images, 4 busy scenes), typed in by script, first image kept, every written element counted at full size.
65 of 97 elements were right (67.0 percent). Portraits and products: 40 of 44. Set-wording images: 15 of 25. The four busy scenes: 10 of 28.
Kitchen scene 6 of 8: more than three apples, a cat under the table where I asked for a dog. Market scene 2 of 7: a crowd instead of two people and no melon being passed. Living room 1 of 6: a green sofa, three striped cushions and no lamp beside it. Beach scene 1 of 7: four umbrellas, tiny chairs, no clear child building a sandcastle.
Added content that changed the scene: invented lines of tiny lettering on the labels, posters and the book cover.
- 17 of 40 prompts. Not run: two portraits, two product shots, one set-wording prompt, two busy scenes, the ten styles and the six character scenes.
Photorealism & Anatomy
3/5Tested- Method
- Eight portrait and product images (P1, P2, P4, P6, F1, F2, F3, F4), one run each, inspected at 100 percent for faults that give a generated image away.
- Images that pass
- 5 of 8
- Hand and face faults
- 1
- Remark
- Scale 3: 13 to 17 of 24 pass. The sample is 8 images, so 5 of 8 (62.5 percent) stands for 15 of 24.
Raw notes
Eight portrait and product images (P1, P2, P4, P6, F1, F2, F3, F4), one run each, inspected at 100 percent for faults that give a generated image away.
Passed (5): the man on the bench, the laughing woman, the woman in the green scarf, the white trainers and the pancakes.
Failed (3): the chef's hand and the knife merge into one and the tomato is not sliced, the olive oil label prints "OlIN OIL" in made-up letters, and the watch dial prints a brand that is not a word.
Hand and face faults 1. Faults in text-like marks on objects 2.
- Eight images in one run, not 24 images in two runs.
Text in Images
1/5Tested- Method
- Five of the six set-wording prompts: the coffee shop sign and board (11 words), the honey jar label (8 words), the night market poster (13 words), the book cover (8 words) and the handwritten fridge note (10 words), 50 words in all.
- Words correct
- 12 of 50
- Images fully correct
- 0 of 5
- Remark
- Scale 1: under 40 percent of words correct, or no readable text. 24 percent.
Raw notes
Five of the six set-wording prompts: the coffee shop sign and board (11 words), the honey jar label (8 words), the night market poster (13 words), the book cover (8 words) and the handwritten fridge note (10 words), 50 words in all.
12 of the 50 words were spelled right, in the place asked for (24 percent). No image was fully correct.
The sign reads "HAROUR BEAAN COFFEE ROASTERRS", the poster headline "SUMER NIGHT MARKET", the honey label "WLD BLOSOM" and the book cover credits "HELE AND MAHASH". The fridge note scrambles the sentence: "PLESE PUT THE" then "MLK REEFGRAOR".
Every image also carried extra lines of letters that read nothing.
- Five of six prompts; the street sign was not run. The model I used is PicLumen's own 1-Lumen model, not GPT-Image-2 at 12 Lumens, which the balance could not afford more than once.
Resolution & Upscaling
3/5Tested- Method
- Every saved image was measured, the frames were read in the app, the product's Download button was used once and the upscaler's price was read.
- Native size
- 1,024 px (1.05 MP)
- Upscale price
- 40 Lumens
- Remark
- Scale 3: 1 megapixel or more native, with no upscaler or a soft one.
Raw notes
Every saved image was measured, the frames were read in the app, the product's Download button was used once and the upscaler's price was read.
All 17 images came out at 1,024 by 1,024 pixels (1.05 megapixels) as WebP in the app. The model offers only 1:1, 5:6 and 6:5. The Download button gave a PNG of the same size.
The upscaler shows "Upscale 40", which is twice the daily balance, and the plans table caps the free plan at 1.5x. 4x, 6x and 8x carry the paid mark. I could not upscale an image, so there is no upscale grade input.
A 1-megapixel native size with no usable free upscaler is the middle mark.
- No upscale was run, so the grade rests on the native size and what the plan allows.
Speed per Image
4/5Tested- Method
- Each of the 17 generations was timed from pressing Generate to the full-size image on screen.
- Median per image
- 10.9 s
- Slowest image
- 14.5 s
- Remark
- Scale 4: a median of 10 to 19.9 seconds.
Raw notes
Each of the 17 generations was timed from pressing Generate to the full-size image on screen.
Median 10.9 seconds, average 11.3, fastest 10.8, slowest 14.5.
The app tagged each image "Fast" and showed "Queued" for a moment before it drew. No generation failed.
The image was ready in about 11 seconds every time, with little spread.
- 17 generations, not 40, on the 1-Lumen model. The 12 to 65 Lumen models were not timed.
Usage Rights, Watermarks & Privacy
2/5Tested- Method
- The terms of service, privacy policy and copyright policy (read 6 November 2026), the plans table and the downloaded file 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, privacy policy and copyright policy (read 6 November 2026), the plans table and the downloaded file 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 (clause 7.1, jointly owned where local law is silent); a paid plan allows commercial use (Lite limited, Standard full); my images stayed in my own account until I chose Post to Community.
Not met (5): the free plan has no commercial licence in the plans table; the free download carries a visible PicLumen mark; the licence in clause 7.4 covers public content for "training AI models" and I found no opt-out; no source for the training data is stated; the file carries no content credentials.
Not tested (2): watermark-free paid downloads, and what the app does when asked for a real person.
- Terms read by me, not a lawyer. No paid plan was bought.
Pricing & Value
5/5Tested- Method
- The free tier was used until I reached the 85 percent floor, and the plans page and its comparison table were read.
- Free images a day, first week
- 20
- Cost per image on Lite
- about $0.010
- Remark
- Scale 5: under $0.02 an image with at least 10 free images a day. That is true for the 1-Lumen model I used and not for the top models. The free images are a welcome offer for seven days, not a standing allowance.
Raw notes
The free tier was used until I reached the 85 percent floor, and the plans page and its comparison table were read.
Free: 20 Lumens a day in my first week only, one image on PicLumen's own model 1 Lumen, so up to 20 images a day; I made 17 and stopped at 3 Lumens left.
The cheapest paid plan is Lite at $7.59 a month, or $6.09 a month paid yearly, with 750 Lumens, about $0.010 an image on the 1-Lumen model.
The top models cost far more: GPT-Image-2 is 12 Lumens, about $0.12 on Lite, and Nano Banana Pro is 65.
- No plan bought; the paid price per image assumes the same 1-Lumen cost.
- The pricing page's Monthly switch can be clicked by hand and I read Lite $7.59, Standard $24.99, Pro $49.99 and Elite $359.99 on it, but the price tracker cannot drive that switch (it is not reachable by its label text), so its monthly figures are the ones PicLumen shows struck through beside the yearly price.
Coverage
Tested
- 17 of the 40 fixed prompts, first image kept: 4 portraits, 4 product shots, 5 set-wording images, 4 busy scenes
- Time and Lumen cost of every image
- The product's own Download button, opened and compared with the stored file
- The model list, the upscaler's price, the plans page and its comparison table
- Terms of service, privacy policy and copyright policy read
Not tested or limited
- The free balance of 20 Lumens paid for 17 images at 1 Lumen each: 17 of the 40 prompts were run, on PicLumen Realistic V2, not the most expensive model
- Style range, consistency and editing were not graded: the balance could not pay for the extra images, and the editing tools carry the paid mark
- Photorealism used one run of 8 images, not two runs of 12 prompts
- The upscaler shows a price of 40 Lumens and the balance was 3, so no image was upscaled
- No plan was bought: watermark-free paid downloads, paid commercial rights 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
- The 12 to 65 Lumen models, especially GPT-Image-2 for words
- Photorealism with 24 images, and the other 23 prompts
- Watermark-free downloads, paid commercial rights and cancellation after buying Lite