Test log
AnimeGenius test data
Raw results for the AnimeGenius 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
- 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: 35 of 49 | 3/5 |
| T05 | Photorealism & Anatomy | Images that pass: 2 of 5 | 2/5 |
| T06 | Text in Images | Words correct: 11 of 11 | 4/5 |
| T07 | Resolution & Upscaling | Native size: 512 px (0.26 MP) | 3/5 |
| T08 | Speed per Image | Median time per image: 14.0 s | 4/5 |
| T09 | Usage Rights, Watermarks & Privacy | Points met: 4 of 8 checked | 2/5 |
| T10 | Pricing & Value | Free images a day: 10 (Realistic) | 5/5 |
| Overall (in the review) | 3.7/5 | ||
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 at 5 credits cost 50 and only 15 of the 100 credits were left under the 85 percent stop.
- 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 at 5 credits cost 60 credits, more than was left.
- Run on a day with credits to spare.
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: the plans page lists the AI image editor under the paid plans, and the credits that were left were under the stop line.
- Run on a plan that includes the editor.
Prompt Adherence
3/5Tested- Method
- 8 of the 40 fixed prompts (2 portraits, the girl with a balloon, the olive oil and the watch, the kitchen and beach scenes, and the coffee shop sign), typed in one at a time, first image kept, every written element counted at full size.
- Elements correct
- 35 of 49
- Busy scenes
- 4 of 15
- Remark
- Scale 3: 65 to 79 percent of elements correct. The sample was 8 prompts and 49 elements, not 40 prompts and about 240.
Raw notes
8 of the 40 fixed prompts (2 portraits, the girl with a balloon, the olive oil and the watch, the kitchen and beach scenes, and the coffee shop sign), typed in one at a time, first image kept, every written element counted at full size.
35 of 49 elements were right (71.4 percent). Portrait, olive oil and sign prompts: 23 of 23. Girl and watch: 8 of 11. The two busy scenes: 4 of 15.
Girl with a balloon 4 of 6: she looks older than eight and holds the balloon at her chest with both hands, not in her right hand. Watch 4 of 5: the face is silver, not black.
The olive oil image added a second bottle the prompt did not ask for (F1); extras are noted, and the counts stay as given.
Kitchen 4 of 8: one apple, not three, and no dog or cat. Beach 0 of 7: many umbrellas in a cluster, no child, no boat.
- 8 of 40 prompts. The style and character groups were not run.
Photorealism & Anatomy
2/5Tested- Method
- Five portrait and product images (P1, P2, P3, F1, F4), one run each, made on the Realistic model and inspected at 100 percent for faults that give a generated image away.
- Images that pass
- 2 of 5
- Faults at 100%
- 3
- Remark
- Scale 2: 7 to 12 of 24 pass. The sample is 5 images, so 2 of 5 (40 percent) stands for about 10 of 24.
Raw notes
Five portrait and product images (P1, P2, P3, F1, F4), one run each, made on the Realistic model and inspected at 100 percent for faults that give a generated image away.
Passed (2): the man on the bench and the laughing woman in a yellow raincoat, with no hand or face faults.
Failed (3): the girl's dress hem is scribbled with red marks and her balloon looks like a lollipop; the olive oil bottle has three corks stacked on top of each other; the watch dial prints made-up letters.
Hand and face faults 0. Faults in objects and printed marks 3.
- Five images in one run, not 24 images in two runs. All five were made on the 5-credit Realistic model.
Text in Images
4/5Tested- Method
- One of the six set-wording prompts: a coffee shop front with a sign ("Harbour Bean Coffee Roasters") and a board ("Open Daily 7 am to 5 pm"), 11 words, made on Z-Image Turbo.
- Words correct
- 11 of 11
- Images fully correct
- 0 of 1
- Remark
- Scale 4: 85 to 94 percent of words, or four images fully correct. All 11 words were right, but one prompt cannot show the five fully correct images a 5 needs, and the image carried invented marks.
Raw notes
One of the six set-wording prompts: a coffee shop front with a sign ("Harbour Bean Coffee Roasters") and a board ("Open Daily 7 am to 5 pm"), 11 words, made on Z-Image Turbo.
Every one of the 11 words was spelled right, in the place asked for.
The image also carried small invented marks: the word "cro." on a panel inside the shop and lines of letters that read nothing on a paper in the door.
- One of six prompts, on the 50-credit model. The Realistic model was not asked for wording.
Resolution & Upscaling
3/5Tested- Method
- Every saved image was measured, the aspect ratios were read in the app, and the high resolution options were read on the settings panel.
- Native size
- 512 px (0.26 MP)
- Largest free size
- 1,024 px (50 credits)
- Remark
- Scale 3: 1 megapixel or more native with no upscaler found. The 5-credit models make 512-pixel images; the free plan reaches 1,024 pixels only on the 50-credit Z-Image Turbo, one image a day.
Raw notes
Every saved image was measured, the aspect ratios were read in the app, and the high resolution options were read on the settings panel.
The seven Realistic images came out at 512 by 512 pixels (0.26 megapixels). The settings panel shows 512 by 768 for the 2:3 ratio. The Z-Image Turbo image came out at 1,024 by 1,024 (1.05 megapixels) for 50 credits.
High Resolution x1.5 and x2, a custom size and batches of four carry a gem icon and are for paid plans. The download was a 512 by 512 PNG of 466 kilobytes, the same picture as the stored file, with no watermark.
- No upscaler was found on the free text-to-image page and none was tested.
- The Z-Image Turbo panel also showed a 2K option with no gem icon that was not tried (pre-T1.png).
Speed per Image
4/5Tested- Method
- Each of the 8 generations was timed from pressing Generate to the full-size image on screen.
- Median time per image
- 14.0 s
- Slowest image
- 22.1 s
- Remark
- Scale 4: a median of 10 to 19.9 seconds.
Raw notes
Each of the 8 generations was timed from pressing Generate to the full-size image on screen.
Median 14.0 seconds, fastest 6.7, slowest 22.1.
The seven Realistic images took 6.7 to 16.5 seconds; the Z-Image Turbo image took 22.1.
No generation failed. The account showed no queue.
- 8 generations, not 40.
- Every generation ran while the app showed 'New User Benefit: Generation speed-up (24h)' with the High Priority / Speed Up toggle switched on (pre-P1.png, 28-claim-offer.png), so a free account after its first day may be slower.
Usage Rights, Watermarks & Privacy
2/5Tested- Method
- The terms of service, content policy and privacy policy (read 6 November 2026), the plans page and the downloaded file were checked against the ten-point list.
- Points met
- 4 of 8 checked
- Remark
- Scale 2: 4 or 5 of 10. 4 of 8 checked is 50 percent, so about 5 of 10.
Raw notes
The terms of service, content policy and privacy policy (read 6 November 2026), the plans page and the downloaded file were checked against the ten-point list. Eight points could be checked on the free plan.
Met (4): the terms say who owns the output (clause 4.1: no claims on ownership or copyright, and created content may be freely used by you); the plans page lists commercial use with the paid plans; the downloaded file had no watermark; images sit in My Artworks and posting is a separate button.
Not met (4): commercial use on the free plan is not stated; the terms and privacy policy say nothing about training on prompts or uploads; no stated source for training data; no content credentials in the file (no provenance markers found in its bytes).
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 allowance was used up to the 85 percent stop, and the plans page and the credit packs were read.
- Free images a day
- 10 (Realistic)
- Cost per Realistic image
- about $0.007
- Remark
- Scale 5: under $0.02 an image with at least 10 free images a day. The 50-credit models cost more.
Raw notes
The free allowance was used up to the 85 percent stop, and the plans page and the credit packs were read.
Free: 50 credits a day, and a Realistic image at the 512 pixel square size costs 5, so 10 images a day. Z-Image Turbo costs 50, one image.
Basic is $9 a month for 8,000 credits, which makes a 5-credit image about 0.6 cents. The smallest one-time pack, 7,000 credits that do not expire, makes it about 0.7 cents. A 50-credit image costs 6 to 7 cents.
Credits are non-refundable (clause 3.5 of the terms).
- No plan bought; annual prices and the paid models' real costs were not read.
- The default output size, 2:3 at 512 by 768, showed 10 credits on the Generate button for the Realistic model and was not used; every image was made at the 512 by 512 square size for 5 credits (30-model-Realistic.png, 2x-settings.png).
Coverage
Tested
- 8 of the 40 fixed prompts, first image kept: 3 portraits, 2 product shots, 2 busy scenes and 1 set-wording image
- Time and credit cost of every image
- The product's own Download button, opened and compared with the stored file
- Terms of service, content policy, privacy policy, storage notice and the plans page read in full
- The locked options on the free plan: premium models, higher resolution, custom size and batches
Not tested or limited
- The free plan's 100 credits that day paid for 8 images: 8 of the 40 prompts were run
- Style range and consistency were not run, and editing was not run because the plans page lists the AI image editor with the paid plans
- Photorealism used one run of 5 images, not two runs of 12 prompts, and text used one of six set-wording prompts
- Premium models (GPT Image 2.5 Flare, Seedream 5, Nano Banana 2.1), high resolution and custom sizes are locked on the free plan and were not tested; the 2K option on the Z-Image Turbo panel showed no gem icon and was not tried
- Every generation ran while the app showed 'New User Benefit: Generation speed-up (24h)' with the High Priority / Speed Up toggle switched on (pre-P1.png, 28-claim-offer.png), so a free account after its first day may be slower. Grade 4 is the measured median.
- No plan was bought: the paid watermark rule, paid commercial use and cancellation were not tested
- Every Realistic image was made at the 512 by 512 square size for 5 credits; the default 2:3 size at 512 by 768 showed 10 credits on the Generate button and was not used
Still to test
- Style range and consistency on a day with credits
- Photorealism with 24 images, and the other prompts
- Image editing on a paid plan
- The premium models, high resolution and the paid watermark rule after buying Basic