Benchmark RadarMEDit-Bench
Evaluates message-driven narrative video editing with long-form videos paired with multiple editing messages and multiple professional edits per message, using temporal alignment metrics and additional annotations for message ambiguity and contextfulness.
- Released
- 2026-07-28
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Addresses the gap in video editing evaluation by accounting for diverse editorial intents, providing a protocol to compare model and human performance on narrative-driven editing, and offering stratification by message difficulty.
Motivation
Video editing is fundamentally message-driven: even from the same source footage, the selected shots change depending on the narrative the editor wishes to convey.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.