Alpha
vs Beta
Same 50 prompts, same reviewers, same battery. One wins on quality, the other on cost. Scores below are from the October 2026 test run.
Alpha if quality is the product. Beta if volume is the product. Alpha took the top blind score in 68% of paired comparisons; Beta delivers 89% of that quality at 10× lower per-image cost. Index scores: Alpha 9.3 · Beta 8.9.
Scorecard
| Dimension (weight) | Alpha | Beta |
|---|---|---|
| output_quality (40%) | 9.6 | 9.1 |
| prompt_adherence (25%) | 9.4 | 8.6 |
| editing_control (20%) | 9.0 | 8.2 |
| value (15%) | 8.1 | 9.7 |
| weighted_total | 9.3 | 8.9 |
Alpha's edge
Subject identity stays stable through successive inpaints — the exact failure mode that cost Beta points in the editing battery. If you iterate on a hero image for hours, this is the difference that matters.
Beta's edge
$0.004 per image against Alpha's $0.04 (published pricing, verified 2026-10). At 10,000 images a month, that is $40 vs $400. For catalogs and A/B creative volume, the math is closed.
Alpha holds #1 and Beta #2 on the main 2026 index. The 0.4-point gap is real. Whether it matters depends entirely on which side of the quality/cost line your work lives.