Video AI Models

Sora 2 offers synchronized audio, 4K quality, and improved physics for cinematic video.

OpenAI's revolutionary video generation model capable of creating realistic and creative scenes from text instructions.

Kling 3.0 offers multi-shot generation, physics-aware motion, native 4K output, and synchronized audio in a unified model.

Kling 2.6 features native audio generation and advanced control over visual consistency and character movement fidelity.

Kling O1 integrates various video tasks into a unified architecture including reference-based generation and keyframe interpolation.

Runway Gen-3 Alpha delivers cinematic-quality visuals with expressive human characters and fine-grained temporal control.

Pika 2.1 introduces high-definition 1080p video generation with Pikadditions for seamless object insertion.

Pika 2.0 added Scene Ingredients feature for integrating user-uploaded images into AI-generated videos.

Pika 1.5 introduced Pikaffects, enabling imaginative transformations like inflating or melting objects.

Luma Dream Machine is a text-to-video model capable of generating realistic motion from user prompts or still images.

Luma Ray3.14 provides native 1080p video generation, 4x faster and 3x more cost-effective with improved motion consistency.

Luma Genie transforms text and images into high-quality 3D assets in minutes.

SeeDance 2.0 is a professional-grade AI video model processing text, images, audio, and video concurrently with cinematic quality.

SeeDance 1.0 focused on transforming static images into fluid, natural-looking videos.

Google Veo 3.1 offers native 4K resolution, improved character consistency, and support for vertical video formats like YouTube Shorts.

Google Veo 3 delivers high-fidelity 8-second clips in 720p or 1080p with integrated audio generation.

Google Veo 2 creates high-quality videos with accurate prompt interpretation and realistic physics simulation.

Hailuo 2.3 offers enhanced visual quality, improved motion coherence, and superior prompt understanding with refined cinematic aesthetics.

Hailuo 02 is a cinematic AI video model producing professional-grade videos with ultra-realistic physics simulations.

MiniMax Video-01 is the foundation model offering multimodal capabilities for video generation.

Mochi 1 is an open-source text-to-video model with 10 billion parameters, offering strong prompt adherence and high-fidelity motion.

LATTE3D generates textured meshes in seconds, acting as a virtual 3D printer.

Haiper 2.5 introduced API integrations before the service was discontinued in early 2025. Now acquired by NetMind.AI.

NIM Microservice is a text-to-3D model from NVIDIA.