01NO OUTPUT YETEvidence required before scoring
Open test slot
- Tool + model
- Not assigned
- Completed runs
- 0 of 3
- Failures
- Must be retained
- Render time
- Awaiting evidence
- Effective cost
- Awaiting evidence
- Output hash
- Awaiting evidence
Same job. Same prompt. Every run counts.
BestVideo.tools will never publish a fastest, best-quality, best-value, or independently-tested badge from provider claims. Each tool must complete three documented attempts under the same task and settings policy.
Make a vertical social clip: “A handheld close-up of a tiny food truck serving miniature mofongo at golden hour, realistic steam, natural camera movement, 9:16.”
02NO OUTPUT YETEvidence required before scoring
Open test slot
- Tool + model
- Not assigned
- Completed runs
- 0 of 3
- Failures
- Must be retained
- Render time
- Awaiting evidence
- Effective cost
- Awaiting evidence
- Output hash
- Awaiting evidence
03NO OUTPUT YETEvidence required before scoring
Open test slot
- Tool + model
- Not assigned
- Completed runs
- 0 of 3
- Failures
- Must be retained
- Render time
- Awaiting evidence
- Effective cost
- Awaiting evidence
- Output hash
- Awaiting evidence
Benchmark protocol
- Record the exact tool, model version, plan, settings, date, prompt, and target duration.
- Run every tool three times. Keep timeouts, moderation blocks, and unusable generations in the record.
- Capture start/end time, credits charged, effective USD cost, and the unedited output file hash.
- Score prompt adherence, motion coherence, visual quality, artifact severity, and creator usability with the published rubric.
- Choose a representative output only after all runs are scored; explain the selection and preserve the others.
Evidence required for a badge
A badge needs a complete run record, retained output evidence, an explicit comparison set, and a current verification date. “Fastest tested” applies only to that dated test set. “Best free option” additionally requires verified plan, card, watermark, and allowance facts.
Current state
No benchmark outputs or scores have been entered. The reusable template is stored in data/benchmark-template.json; the empty state above is intentional.