xAI推出视频生成模型grok-imagine-video-1.5-preview，目前在Artificial Analysis Video Arena的Image to Video (With Audio)排行榜中排名第二，仅次于字节跳动Seedance 2.0。该模型支持图像转视频并原生生成音频，最长可生成15秒视频。在无音频排行榜中位列第三，紧随Seedance 2.0和自家的grok-imagine-video。模型定价为每分钟视频$8.40，现已通过xAI API提供，并将逐步在Grok app和X上线。
Grok debuts grok-imagine-video-1.5-preview, achieving #2 in Image to Video (With Audio) in the Artificial Analysis Video Arena, behind only ByteDance's Seedance 2.0!
grok-imagine-video-1.5-preview is @xAI's latest video generation model, currently supporting only Image to Video with native audio, and durations up to 15s. It ranks #2 in the Image to Video (With Audio) Leaderboard, trailing only ByteDance's Seedance 2.0. In the Without Audio Leaderboard it places #3, behind Seedance 2.0 and xAI's own grok-imagine-video, which it performs very closely to.
grok-imagine-video-1.5-preview costs $8.40 per minute of generated video, and is available now via xAI's API, with a broader rollout across the Grok app and X in progress.
Congratulations to @xAI and @elonmusk on the release!
See below for comparisons between grok-imagine-video-1.5-preview and other leading models in the Artificial Analysis Video Arena 🧵
来源：Artificial Analysis · x.com