The evaluation, conducted by San Francisco-based Physion Labs, tested seven agents across 700 videos and 100 distinct prompts. Agent One dominated the field by leading all three core quality dimensions: Narrative Coherence, Cinematic Language, and Production Quality. It secured first place on 12 of the 16 metrics tracked, never dropping below second in any category. Most notably, the system achieved a 97.8 score in Narrative Alignment, the highest individual result recorded during the testing process.

Physion Labs, led by CEO Evelyn Qin Zhang and CTO Bing Shuai, utilized human annotators to score the outputs, with a third-party adjudicator resolving disagreements to ensure impartiality. The benchmark results indicate a statistically significant lead for Invideo, with a p-value of 0.046 over the second-place contender. Sanket Shah, founder and CEO of Invideo, attributed the performance to technical strategy rather than capital scale, noting that the agent is designed to maintain narrative continuity for creative directors.