apiex选择模型后可复制配置给 AI，也可以快速测试当前连接延迟。
Alibaba
Extends a single starting image into a full shot, preserving its composition and style, and outputs 3–15 second clips at up to 1080p. Suited for turning still assets into motion.
Alibaba
Takes a set of reference images to lock character, scene, and style, holding the subject consistent across shots. Outputs 3–15 second clips at up to 1080p.
Alibaba
Generates 3–15 second clips at up to 1080p from a written description, across a range of aspect ratios. Suited for taking a script straight to creative content and short-form social video.
Alibaba
Smoother motion than 1.0 across camera moves and body movement. Extends a single starting image into a full shot, outputting 3–15 second clips at up to 1080p.
Alibaba
Holds characters more consistent across frames than 1.0. Takes a set of reference images to lock character, scene, and style, outputting 3–15 second clips at up to 1080p.
Alibaba
Stronger prompt adherence than 1.0 on complex instructions. Generates 3–15 second clips at up to 1080p from a written description, suited for producing creative content directly from a script.
Alibaba
Generate 5-10 second, 480p/720p/1080p video from text prompts with native synced audio track.
Alibaba
Alibaba Tongyi Wanxiang Wan 2.6 image-to-video model animates a reference image into 720p/1080p clips with synchronized audio, preserving subject consistency with cinematic motion.
Alibaba
Speed-optimized Wan 2.6 image-to-video variant with optional audio output, generating 720p/1080p clips quickly and cost-effectively for high-throughput image animation.
Alibaba
Alibaba Tongyi Wanxiang Wan 2.6 reference-to-video model replicates the actions, effects, and camera movement of a reference video to produce up to 10 second 720p/1080p clips with synchronized audio.
Alibaba
Generate 720p/1080p video from text prompts and come with synchronized audio tracks. The picture has movie-level aesthetics and complex motion performance, suitable for creative video creation.
Alibaba
Speed-optimized Wan 2.6 reference-to-video variant with optional audio output, replicating reference-video motion into up to 10 second 720p/1080p clips for fast, cost-effective production.
bytedance
Standard tier of the Seedance family. Supports text-to-video, image-to-video, and reference-to-video, with strong character, style, and camera consistency. Built for ads and short-form creative.
bytedance
Lightweight Seedance 2.0 variant tuned for speed and cost rather than peak fidelity, with text-to-video, first/last frame control, and multi-reference inputs.
bytedance
A video generation model built for long-form storytelling, with first-frame and first-and-last-frame control, up to 50 multimodal reference assets, and optional audio generation. It also supports video editing and extension.
kling
The oldest generation still available, limited to image-to-video with first- and last-frame control and no multi-image reference. Suited for basic, cost-sensitive clip generation.
kling
Kling 1.6 supports text, first or last frame, and multi-reference-image video generation.
kling
Flagship of the 2.0 line, tuned for instruction following and cinematic aesthetics, with stylized effect presets. Outputs 1080p only.
kling
Kling 2.1 image-to-video generation supports first and last frame control.
kling
Quality-oriented tier of the 2.1 line, tuned for motion performance and semantic responsiveness. Supports text-to-video and image-to-video, suited for single-shot clips.
kling
Cost-efficient tier in the 2.5 generation, with improved prompt adherence and physics under large-amplitude motion. Outputs silent 5- or 10-second clips.
kling
First version to add native audio, producing visuals, voiceover, sound effects, and ambience in one pass. Suited for finished clips up to 10 seconds.
kling
Core model of the 3.0 line, with first- and last-frame control, multi-shot storytelling, and native audio produced in the same pass. Suited for finished short-form work.
kling
Fully multimodal variant of the 3.0 line, accepting text, image, and video references in a single generation, with voice-driven characters and native audio.
kling
Speed-and-cost tier of the 3.0 line, covering text-to-video and first-frame image-to-video with multi-shot output. Suited for high-volume production.
kling
Cinematic-oriented model with first- and last-frame control and video reference input, including edits driven by an existing clip. Suited for narrative shot sequences.
minimax
It supports Vincentian and Tusheng videos, can output 6-10 seconds, 512p/768p/1080p multi-level image quality, supports camera movement command control, and flexibly balances image quality and cost.
minimax
It supports Vincentian and Tusheng videos, can output 6-10 seconds, 768p/1080p images, supports fine camera movement command control, has smooth movement and stable subject, and is suitable for short video creative content.
minimax
Speed and cost optimized variant of MiniMax Hailuo 2.3 for rapid iteration and bulk production, generating 6 to 10 second 768p/1080p clips from text or image input with camera-movement control.
minimax
A lightweight open-weight video generation model focused on instruction-guided editing, text and brand rendering, and video-to-video motion transfer, with native audiovisual output. Built for advertising, e-commerce, and interface design workflows.
OpenAI
It can generate detailed, dynamic 720p videos from text or reference images with synchronized audio tracks. It has excellent understanding of three-dimensional space, motion and lens continuity, and is suitable for the creation of 16-20 second cinematic short films.
xAI
xAI Grok Imagine Video 1.5 generates video from text, a first-frame image, or up to three reference images.