Artificial Analysis launches Code2Video-AA, a benchmark of marketing videos made in code

Artificial Analysis introduced Code2Video-AA, its own implementation of HeyGen's Code2Video Bench, built in partnership with HeyGen. Models get 114 product and brand marketing briefs and must answer each with a HyperFrames HTML composition in a single response, without tools; the composition is rendered to video with no network access. A judge model (Gemini 4 Argon, high) compares pairs of videos on engagement, prompt intent, composition, temporal consistency and craft, and the five axis Elo ratings are averaged into the headline score. More than 35,000 verdicts rank 27 models. The results are Artificial Analysis's own.

Results it reports:

• Claude Opus 5.5 (max with fallback) leads at 1536 Elo, 93 points ahead of the next models, and tops all five axes, with its largest lead on craft. It costs $3.08 per task.
• Claude Sonnet 5.5 (1443) and GPT-6 Astra (1427) are statistically tied for second.
• GPT-6.1 Sol (max) is fourth at 1369 and $0.31 per task, about a tenth of Opus 5.5's cost. Cost per task across models ranges from $0.01 to $3.89, and eight models are on the Elo-against-cost Pareto frontier.
• Open-weights models trail by at least 483 points: DeepSeek V4.1 Flash (max, 1053), MiMo-V2.6-Flash (1052) and MiMo-V2.6-Pro (1052) are the top ones, at $0.01 to $0.08 per task.