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.
More from Companies & People
- As AI Spending Soars, Can Chinese Tech Giants Deliver Long-Term Profits? — pstAsiatech · 2026-07-30
- KDD 2026 Preview: LLM Failure Modes in Multi-Tool Long-Form Analysis — 0xsachi · 2026-07-30
- CNBC Interviews OpenAI CFO: Full Pivot to For-Profit and Restructuring — kimmonismus · 2026-07-30
- Carrier Revolt Bundles ChatGPT Go, Offering Up to 12 Months Free — testingcatalog · 2026-07-30
- Revolut Partners with OpenAI to Offer Free ChatGPT Plus Subscription to Customers — testingcatalog · 2026-07-30
- The New AI Marketing Trend: Why Companies are Choosing Long-Form Video Platforms — lxfater · 2026-07-30