Runway Gen-4.5 tops video benchmarks with superior physics
The "David" model beats Google and OpenAI in blind tests, proving specialized labs can still compete.
Disrupting the hierarchy
Runway, the AI video startup, claimed the top spot on the prestigious "Video Arena" leaderboard on Tuesday, disrupting the dominance of tech giants. Its latest model, Gen-4.5 (codenamed "David"), achieved an Elo rating of 1247 in blind A/B testing, displacing Google’s Veo 3 and pushing OpenAI’s Sora 2 Pro to a distant ranking. This victory is significant because the Video Arena relies on crowdsourced human preference, where users vote on video quality without knowing which model created it.
Physical intelligence vs. dream logic
The primary differentiator for Gen-4.5 is its grasp of "physical intelligence." Previous …
Archive Access
This article is older than 24 hours. Create a free account to access our 7-day archive.