Grok Imagine 1.5 Brings Image-to-Video at 720p to xAI API

SAN FRANCISCO — xAI pushed its visual AI capabilities into new territory on June 3, announcing Grok Imagine 1.5 Preview — an update that adds image-to-video generation to the platform and delivers higher-quality still images for developers and enterprises building on the Grok API.

The launch comes as the company's combined compute and model portfolio continues to expand at a rapid pace, adding another dimension to the multimodal assistant Elon Musk has positioned as the backbone of xAI's consumer and developer ecosystem.

Image-to-Video at 720p

The headline feature of Grok Imagine 1.5 is grok-imagine-video-1.5-preview, an image-to-video model that converts still images into cinematic short clips at resolutions up to 720p, driven by text prompts. Developers pass an existing image alongside a descriptive prompt and receive a fluid video output that brings the scene to life.

The model is available immediately through the xAI API at x.ai/api/imagine. xAI positioned the capability as an entry point for product demos, ad creation, and social content — use cases that previously required expensive dedicated video generation services.

Shortly after the launch, Elon Musk demonstrated the capability publicly by posting an AI-generated trailer for the Iliad, the ancient Greek epic, created entirely with Grok Imagine 1.5. The clip drew widespread attention and underscored the creative range the model supports.

Quality Mode Raises the Bar for Still Images

Alongside the video model, xAI introduced a Quality Mode for the Grok Imagine image generation and editing API. The mode runs four high-quality images per generation request rather than the standard single output, delivering improvements across detail, lighting, shadow rendering, and text accuracy within images.

The endpoint is identified as grok-imagine-image-quality and is aimed at enterprise production workflows including product photography, marketing assets, ad variants, and branded visuals. xAI described it as designed for cases where image fidelity matters more than generation speed.

Growing the API Ecosystem

The Imagine 1.5 announcement follows a series of rapid releases across the Grok platform. Grok Voice went live on June 4, bringing spoken conversation to all users. Grok Build 0.1, the dedicated coding model with a 256,000-token context window, launched in public beta in late May. And the Grok API recently added worktrees support, enabling developers to work across multiple branches of a repository simultaneously within a single Grok session.

Together, these updates reflect xAI's strategy of expanding Grok from a consumer chatbot into a full-stack developer platform — one that handles conversation, code, images, video, and voice inside a unified API.

What Comes Next

xAI has indicated that Grok V9 Medium, trained at 1.5 trillion parameters, is approaching a public release in mid-June 2026. The model is expected to deliver sharper reasoning, stronger coding output, and more reliable performance on complex queries compared to the current production model.

With image-to-video now live and a major model upgrade on the horizon, the Grok ecosystem is moving quickly. For developers already building on the xAI API, the timing could not be better — each new capability adds a layer of functionality that would otherwise require integrating a separate third-party service.