InVideo vs. Runway: Do AI Video Benchmarks Measure Models or Orchestration?

taherdhanera · x · 2026-07-30

This debate stems from InVideo's 'Agent One' benchmark, which calls external models like Seedance, Veo, and Kling during evaluation. Some sarcastically suggested this asks Indian companies to 'win quietly and apologise.'

The author counters that if the benchmark evaluates external models, it measures the quality of the complete orchestration system, not InVideo's foundational video model. The author emphasizes that Runway, as a pioneer, invested heavily in proprietary foundational video and world models, building a deep moat across research, professional tools, film funding, and creator ecosystems—something companies relying primarily on external providers cannot match.

Original post →

More from Companies & People

Companies & People channel →