Grok Imagine Video v1.5 Image-to-Video API by xAI

1. Introduction

Grok Imagine Video V1.5 is a frontier-tier image-to-video generation model developed by xAI that animates static images into short clips of up to 15 seconds with natively generated, synchronized audio — including dialogue, lip-sync, sound effects, and ambient music — produced in a single inference pass.

This README applies to the following API model identifier:

• xai/grok-imagine-video-v1.5/image-to-video

Released in preview around late May 2026, Grok Imagine Video V1.5 debuted at the top of the Artificial Analysis Video Arena Image-to-Video leaderboard with a 1404 ±6 Elo rating, surpassing ByteDance Seedance 2.0 and other established competitors. Built on xAI's Aurora engine — an autoregressive mixture-of-experts (MoE) network that jointly models text, image, video, and audio tokens — the model represents a departure from the diffusion-transformer paradigm used by Sora and Veo, enabling tightly coupled audiovisual generation with competitive cost and latency characteristics.

2. Key Features

• Native Synchronized Audio Generation: Audio (dialogue, lip-sync, SFX, ambient sound, music) is generated jointly with video tokens in a single inference pass rather than dubbed in post-processing. This produces event-aligned sound effects and natural lip-sync without requiring separate audio pipelines.

• Aurora Autoregressive MoE Architecture: Unlike diffusion-transformer competitors, V1.5 uses an autoregressive mixture-of-experts network trained to predict next tokens from interleaved multimodal data. This unified token-space approach is what enables single-pass audio-video coherence.

• Granular Duration Control (1–15 seconds): Clips can be requested at any integer second from 1 to 15, supporting precise targeting for short-form formats. V1.5 extends the prior 10-second limit by 50% while maintaining temporal coherence across the longer window.

• Improved Physics and Photorealism: V1.5 introduces measurable gains in cloth dynamics, water simulation, hair motion, and object interaction. Subject deformation in high-motion scenes is reduced relative to V1.0, with sharper micro-expressions and improved translucent/glass material rendering.

• Fast Inference: A 5-second 720p clip generates in approximately 20–30 seconds end-to-end — roughly 2–3× faster than Seedance 2.0.

• Broad Format Support: The model accepts JPG, JPEG, PNG, WEBP, GIF, and AVIF input images and outputs H.264 MP4 at 24 FPS across seven aspect ratios (1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3), at 480p or 720p (1280×704) resolution.

• Extend Video Chaining: Optimized clip extension allows users to chain segments into longer multi-shot narratives, with V1.5 improving continuity between extension boundaries relative to V1.0.

3. Model Architecture & Technical Details

xai/grok-imagine-video-v1.5/image-to-video is built on xAI's Aurora engine, an autoregressive mixture-of-experts network that predicts next tokens across an interleaved sequence of text, image, video, and audio modalities. This is architecturally distinct from the diffusion-transformer designs used by OpenAI Sora and Google Veo, and is the mechanism by which V1.5 produces joint video+audio output in a single forward pass rather than chaining separate generative models.

Key infrastructure and lineage points:

• Training infrastructure: Trained on xAI's Colossus 2 supercomputer, a ~2 GW, ~555,000 NVIDIA GPU facility — the largest known single-site AI training cluster.
• R&D lineage: The video pipeline incorporates technology from Hotshot, a video generation startup acquired by xAI in March 2025.
• Aurora foundation: The underlying Aurora image model was first released on December 9, 2024, with video capability progressively layered on top through Imagine 0.9 (October 2025), Imagine 1.0 (February 2026), multi-image and extension support (March 2026), and the V1.5 preview (May 2026).
• Joint token modeling: Because audio and video tokens are produced in the same autoregressive stream, lip-sync and event-aligned SFX emerge from the model rather than from separate alignment models.

xAI has not published a technical report, parameter count, training-data disclosure, or formal model card for V1.5, so finer architectural details (expert count, context length, tokenizer design) are not publicly documented.

4. Performance Highlights

xai/grok-imagine-video-v1.5/image-to-video debuted at #1 on the Artificial Analysis Video Arena Image-to-Video leaderboard with an Elo rating of 1404 ±6, displacing ByteDance Seedance 2.0 from the top spot.

Comparative positioning across leading image-to-video and video-with-audio systems:

Qualitative performance characteristics:

• Image-to-video coherence: Currently the top-ranked model on the Artificial Analysis I2V arena, particularly strong on photorealistic portrait animation, micro-expressions, and translucent material rendering.
• Audio quality: Sharper lip-sync and cleaner voice rendering than V1.0; still trails Veo 3.1 on lip-sync precision for dense dialogue.
• Throughput: Approximately 2–3× faster inference than Seedance 2.0 at comparable resolution.
• Scale of adoption: V1.0 reportedly generated 1.245 billion videos in its first 30 days of availability, indicating substantial production-scale deployment.
• Known weaknesses: Physics fidelity in combat and collision scenes lags top competitors; 720p output cap places it below 1080p-capable rivals for high-resolution delivery.

5. Use Cases

• Short-Form Social Video: Vertical (9:16) and square outputs at 1–15 seconds map directly to TikTok, Instagram Reels, YouTube Shorts, and X clip formats, with native audio eliminating the need for separate sound design.

• Marketing and Advertising Creative: Rapid generation of product visuals, brand teasers, and ad concepts makes the model suitable for high-volume creative iteration and A/B testing of motion concepts.

• Image Animation: Static portraits, posters, illustrations, and product photography can be animated with motion and synchronized audio, enabling reanimation of existing brand and editorial assets.

• Concept Visualization and Pre-Visualization: Fast 20–30 second inference per 5-second clip supports rapid concept testing for filmmakers, designers, and creative directors who need to evaluate motion and audio direction before committing to full production.

• Multi-Shot Narratives via Extend Video: The optimized extension pipeline supports chaining clips into longer sequences, suitable for short narrative pieces, episodic memes, and serialized social content.

• Game and Interactive Asset Pipelines: The text → image → animated video flow integrates into game development and interactive media workflows for cinematics, character idle/action loops, and trailer footage.

• Entertainment and Viral Content: Native distribution through Grok on X, combined with low cost and granular duration control, supports meme, parody, and viral content generation directly inside the X ecosystem.

The model is less well-suited to long-form storytelling, structured brand-consistent campaigns requiring fine-tuning, and applications requiring 1080p or higher output resolution.