Artificial Analysis Launches AA-Video-I2V v2.1 Image-to-Video Benchmark

Artificial Analysis launched AA-Video-I2V v2.1 and AA-Video-I2V-Silent v2.1, benchmarks evaluating image-to-video models at up to 1080p across 10 use cases, 10 capabilities and diverse visual styles. The benchmarks draw on more than 58,000 and 44,000 human preference votes respectively, with only 5% of prompts public and rotated monthly. Dreamina Seedance 2.5 ranks first overall, statistically tied with Seedance 2.0 and the MiniMax H3 family, while MiniMax H3 Max tops the Silent leaderboard.
Event version 1.
Key insights
Benchmark scope and structure
AA-Video-I2V v2.1 evaluates models at up to 1080p using prompts that pair a 1080p starting image with a written prompt. Prompts are tagged by use case, capability and visual style, and the overall ranking samples evenly across use cases and capabilities to reflect versatility.
Vote volume and contamination resistance
AA-Video-I2V v2.1 includes more than 58,000 human preference votes from private US/UK evaluators over 1,000 prompts. AA-Video-I2V-Silent v2.1 includes more than 44,000 votes over 500 prompts, and only 5% of prompts are public, rotated monthly.
Dreamina Seedance 2.5 leads overall
Dreamina Seedance 2.5 ranks #1 overall, statistically tied with Seedance 2.0 and the MiniMax H3 family, and leads 11 category boards including Human Anatomy, Camera Control and Retail & Ecommerce. At $34.12/min it is the most expensive model in the top 10.
MiniMax H3 pricing and category leads
MiniMax H3 (768p), an open weights model, ranks #3 at $4.80/min, about one-seventh the price of Seedance 2.5, and is #1 on Text Rendering. MiniMax H3 Max, post-trained by fal, ranks #4 with audio and #1 on AA-Video-I2V-Silent v2.1.
Silent leaderboard and other category leaders
Bach 1.0 Pro from Video Rebirth ranks #3 on AA-Video-I2V-Silent v2.1, statistically tied for #1, at $7.20/min. Dreamina Seedance 2.0 ranks #2 overall and #1 on Audio Synchronization and Lighting & Materials.