BestAIImageGen

Benchmark results

One prompt pack, multiple tools. Notes below stay honest about free-tier limits-screenshots will be added as each tool run is completed.

Test protocol

Status: Qualitative scorecard below is live. Visual cells update as screenshot runs complete. Rankings on the homepage stay the quick pick.

Preliminary scorecard

0–5 per axis from the published checklist. Not a lab measurement-honest editorial scores until the image grid is filled.

Tool Prompt follow Hands / detail Text stress Free path Export Total /25
Midjourney 4 5 3 2 4 18
DALL-E / ChatGPT 5 4 4 3 4 20
Adobe Firefly 4 3 3 3 5 18
Canva AI 3 3 3 4 5 18
Perchance 3 2 2 5 3 15

Last editorial pass: 2026-03. Re-score when screenshot rows below are filled. Download: benchmark-scorecard-v1.csv.

Comparison grid

Each cell = prompt × tool. Drop WebP screenshots into /benchmark-results/img/ with descriptive alt text.

Prompt Midjourney DALL-E / ChatGPT Firefly Canva Perchance
1 Portrait Pending capture Pending capture Pending capture Pending capture Pending capture
2 Product Pending capture Pending capture Pending capture Pending capture Pending capture
3 Landscape Pending capture Pending capture Pending capture Pending capture Pending capture
4 Text stress Pending capture Pending capture Pending capture Pending capture Pending capture
5 Stylized Pending capture Pending capture Pending capture Pending capture Pending capture
6 Flat lay Pending capture Pending capture Pending capture Pending capture Pending capture

Preliminary qualitative notes

Head-to-heads: all comparisons.

FAQ

Why are screenshots pending?

We publish the protocol first so others can reproduce. Visual cells update as each tool is run on the same pack.

Can I reuse the prompts?

Yes. Copy from benchmark-prompts and link back if you publish a comparison.