Playcut Benchmarks
A versioned, reproducible public record of what AI actor and AI UGC platforms actually deliver — measured the same way across every vendor, published whether the result flatters us or not.
Playcut builds one of the platforms measured here
That is a real conflict of interest, and the answer is structure rather than a claim of neutrality. The protocol is fixed before any scoring run, Playcut is labeled as the vendor on every page, and the results where Playcut loses are published alongside the rest.
Why this program exists
Buying decisions in this category are made almost entirely on marketing claims. Vendors publish capability lists and headline resolutions; almost none publish a number you can compare against a rival, and several do not publish a price at all. The result is a market where the loudest claim wins rather than the best product.
These pages are the correction. Every platform is checked on the same columns on the same date, using the same definitions. Where a vendor's own page is the source, it is linked. Where a claim cannot be independently confirmed, it is marked as a claim rather than promoted to a fact.
The intended outcome is that a buyer can finish a decision here — and that other publications cite this data rather than recycling vendor copy.
The three rules
Honest numbers
A published figure is a real, sourced value or it is absent. Nothing is estimated to make a table look complete.
Vendor disclosure
Playcut is marked as the vendor wherever it appears, and losses are published as readily as wins.
Versioned forever
Releases carry a version and a compile date, and keep the same URLs, so a citation stays checkable.
Published releases
Three assets are live today. Each is maintained on its own cadence and keeps its URL across versions.
What is published now, and what is not
Published — verified facts
- • Capability comparison across 12 platforms
- • Entry price, metering model, free tier, watermark policy
- • Language coverage, maximum export resolution, public API
- • Reusable cross-surface actor identity
- • Pricing-transparency audit
Not yet published — lab phase
- • Identity consistency (ArcFace) and cross-surface portability
- • Lip-sync (SyncNet) and script fidelity (Whisper WER)
- • Artifact rate and usable-take rate
- • Cost per usable video, generation latency
- • Voice naturalness and consistency
This split is deliberate and worth stating plainly: the current release contains no instrument-measured scores. Do not cite these pages for realism, quality or lip-sync rankings, because they do not contain them. Those metrics publish with their raw data once the lab run completes under the same protocol for every platform, Playcut included.
How to cite this data
Cite the release version and the compile date, not just the page. Figures change between releases, and a citation without a version cannot be checked later.
Corrections are welcome and are the fastest way to improve a release. If a figure here is wrong or out of date, write to contact@playcut.ai with the vendor's own page as evidence and it will be checked against the source.
Frequently asked questions
What is the Playcut benchmark program? +
A public, versioned record of what AI actor and AI UGC platforms actually deliver, measured the same way across every vendor. It exists because the category's buying decisions are made almost entirely on marketing claims — vendors publish capability lists, not comparable numbers. Releases keep their URLs permanently so a citation made today still resolves in two years.
Playcut is one of the platforms tested — how is that not a conflict of interest? +
It is a conflict of interest, and the answer is structure rather than a claim of neutrality. The protocol is fixed before any scoring run, Playcut is labeled as the vendor on every page, results where Playcut loses are published, and every figure is reproducible from the published method. You are invited to check the work rather than trust the source.
What does the current release actually contain? +
Live-verified capability, pricing and policy data only, checked 2026-07-17 against each vendor's own pricing page and product documentation. It contains no instrument-measured scores. Identity consistency, lip-sync, artifact rate, cost per usable video and latency are all deferred to a later release, after the lab run completes under the same protocol for every platform.
Can I cite or reuse this data? +
Yes. Cite the release version and the compile date — for example "Playcut AI Actor Benchmark v0.9, compiled 2026-07-17" — because figures change between releases and a citation without a version cannot be checked. Raw data and scripts publish alongside the lab release.
How often does it update? +
The pricing index data was verified 2026-07-17 and a monthly re-verification sweep is scheduled from 2026-09-01; we will describe it as an established cadence only once it has actually held for two consecutive months. The actor benchmark ships quarterly versioned releases at the same URLs. Price changes and new releases are the moments worth watching — a diffable price history across the category is data nobody else is keeping.