L'article ci-dessous est publié en anglais. Chaque chiffre qu'il contient est le chiffre publié par l'éditeur lui-même, avec la date à laquelle nous l'avons lu.
What separates the tools in this category
Pricing model, more than output quality, is what decides which of these a team can actually adopt. A credit-based web app and a per-second API are not competing products even when they generate similar footage: one is a subscription a marketing team can expense, the other is a line item that scales with usage and needs an engineer to wire up.
The second dividing line is what the tool is for. Runway, Pika, Luma and Veo generate footage from a prompt. Descript is an editor rather than a generator: you edit the transcript and the video follows. Comparing them on "video quality" flattens a distinction that decides the purchase.
Tools that turn a script into a presenter reading it are filed under AI Avatars rather than here, because they are metered in minutes of finished video and bought by different teams. They appear alongside these tools in our video ranking, which ranks all seven on what the free tier grants, names the one allowance it could not read, and says out loud that the units do not convert.
What we publish and what we do not
Every figure on the profiles in this category is the vendor's own published number, with the URL and the date we read it. Where a pricing card printed two contradictory monthly figures we left the field empty rather than choose one, and where a vendor's site refused our request entirely we said so instead of inferring an absence.