TubeLab · Video scripts
TubeLab · Writes for different video channels, comparing story organization, style, pacing, and clarity.
What it measures and how
Uses a common TubeLab Scriptwriter, 12 tasks, and six quality dimensions. Three judges do not judge their own provider's models; failed tasks retain penalties. Cost and speed are excluded from ability scores.
How this evidence is used
Under review: stable raw data, actual evaluation dates, and reuse terms are pending. Display pages do not substitute for verifiable results.
Evaluation limits & attribution
Results depend on Scriptwriter. OpenAI uses its lowest reasoning tier, so results do not generalize to other tiers. Video narration is not screenwriting.
Data license: Stable exports and public-aggregation permission are unconfirmed.
Results are published by TubeLab. Original scores and the NexGateHub consensus index use different scales and cannot be added directly.