What vendors publish about picture quality

Nothing on this criterion falls below three, and six entries reach five. That flatness is the finding: picture quality is where the field has converged, which makes it the least useful column for telling these tools apart. As of 2026-09-12.

Inclusion rule. Tools rated on this rubric: a public English product page addressing at least four of the seven criteria. A tool the vendor documents too thinly to rate is left out rather than rated low. Order. Descending on this criterion, then alphabetical; unrated last.

1A column that no longer separates anything

When six of sixteen sit at the top and none sit at the bottom, the column has stopped doing the work a rating is for. Every vendor here names a model, publishes resolutions and, mostly, a maximum clip length. The differences that remain are real but do not show up in what is published.

That is worth saying plainly on a page that ranks things: the ranking below is accurate and nearly useless as a shortlist. The columns that separate tools are elsewhere.

2What the rating actually measures here

Not how good the output looks. It measures whether the vendor names the model, states resolutions and states a maximum length — three facts a production needs before it can plan. Output quality is a judgement this register does not make and could not make from published pages.

The distinction matters because a reader could take a five as an aesthetic verdict. It is not one. It means the vendor told you what you are buying.

3Where the remaining spread comes from

Entries at three usually publish a resolution and stop: no maximum length, or no named model behind the tier. That gap is small on a rate card and large in a schedule, because clip length determines how many generations an episode costs.

The published lengths themselves are on the clip limits page, and the rates they attach to are on the ledger.

4The three facts behind every rating here

A named model, the resolutions it runs at, and the longest single clip. Each is checkable on a vendor page in under a minute, and each changes what a production can plan. A tool publishing all three can be budgeted before anyone signs up; one publishing a resolution alone cannot.

Where a vendor names a model but not a version, the entry sits lower. A version is what makes output reproducible six months later, and without it the same prompt on the same tier is not the same request.

5Reading this column against the price

Because quality is flat and price is not, price differences on this table are rarely buying picture quality. They are buying resolution tiers, speed, and the surrounding steps. A reader treating a higher rate as a proxy for a better picture will overpay on this evidence.

Ratings describe what each vendor publishes, read 2026-09-12, and are given under the rule on how scoring works. All seven criteria together are on the ratings table. Next: Sound, Assembly.