Artificial Analysis Launches Code2Video-AA Benchmark
The new leaderboard ranks 27 AI models based on their ability to generate video animations from design briefs.
Updated on Oct. 26, 2026 in Artificial Intelligence
Live Poll
Would you use AI models to create marketing videos for your business or brand?
Artificial Analysis has released the Code2Video-AA benchmark, which evaluates how effectively various AI models can create videos. The assessment uses blind pairwise comparisons to assign Elo ratings based on 114 specific design briefs.
Why it matters
This benchmark provides a standardized metric for evaluating the video generation capabilities of generative AI. By testing models against consistent design requirements, it offers a clearer view of performance across the rapidly evolving AI sector.
Claude Opus 5.5 leads the field with an Elo rating of 1536, followed by Claude Sonnet 5.5 at 1443 and GPT-6 Astra at 1427. Task costs for the 27 tested models range from $0.01 to $3.89 per output, with GPT-6.1 Sol costing approximately $0.31.
The players
Artificial Analysis
An organization that specializes in testing and benchmarking artificial intelligence models to provide performance metrics.
The details
Models in the study are required to generate videos by producing HTML compositions that utilize the GSAP JavaScript library for animation. Evaluators then perform blind pairwise comparisons of these rendered videos to determine the rankings.
Timeline
October 26, 2026: Artificial Analysis launched the Code2Video-AA benchmark.
The Tech Race
The introduction of the Code2Video-AA benchmark marks a transition from subjective quality assessment to standardized quantitative evaluation in AI video generation. This framework replaces anecdotal testing with a structured leaderboard, intensifying the performance competition among major AI developers.
Developers and users can use this leaderboard to identify which models provide the best animation results relative to their task costs. Understanding these price-to-performance metrics helps individuals select the most efficient AI tools for specific design workflows.
The takeaway
Standardized benchmarks are increasingly vital as AI video generation moves toward professional and commercial applications. Users should monitor these evolving Elo scores to ensure they are utilizing models that match their specific budgetary and creative needs.
Further reading
For more on how new models are changing the field, visit Artificial Intelligence.
Source note: This article includes information reported by Crypto Briefing.
Live Poll
Would you use AI models to create marketing videos for your business or brand?
