xAI went from having no video product in August 2025 to claiming the top spot on AI video benchmarks by January 2026, and its latest version 1.5 has extended that lead even further. The model, powered by an architecture called Aurora, generated a 40-second trailer for Homer's Iliad that pulled 18.4 million views when Elon Musk posted it on X on June 3, 2026. Version 1.5 arrived at an Elo rating of 1404 on the Artificial Analysis Video Arena leaderboard, a 52-point jump over version 1.0, placing it above competitors including Runway Gen-4.5, Sora 2 Pro, and Google Veo.

What Makes Grok Imagine 1.5 Different From Earlier AI Video Models?

The engine underneath Grok Imagine 1.5 is Aurora, an autoregressive architecture trained on 110,000 NVIDIA GB200 GPUs. Autoregressive means the model generates each frame based on what came before it, rather than producing all frames at once and stitching them together. This approach helps maintain coherence across longer sequences. Earlier AI video models generated frames more independently, which is why motion felt jerky past four or five seconds. Grok Imagine 1.5 runs to 15 seconds at 24 frames per second, and the coherence mostly holds.

One standout feature is native audio generation, which synchronizes dialogue, ambient sound, effects, and music in the same pass as the video. The Iliad trailer demonstrated this capability across scene cuts without requiring separate audio tools or manual sync work. Six months earlier, pulling off that level of audio-visual coordination would have required three separate tools and significant manual labor.

How Many Ways Can You Use Grok Imagine 1.5?

The model supports six distinct generation modes, though one of them has a critical limitation that catches many users off guard:

• Image-to-Video: Upload a still image and describe the motion; Aurora animates outward from that frame, preserving lighting, composition, and subject identity. This is where the model is strongest and most consistent.
• Text-to-Video: Build a scene from a written prompt only, more generative and less predictable than image-to-video. Important caveat: the API version does not support text-to-video; it accepts image input only. Text-to-video works through grok.com and the Grok app.
• Video Extension: Continue a clip from its last frame to build sequences longer than 15 seconds. Version 1.5 handles the joins better than 1.0, though face drift and lighting inconsistencies still accumulate after several chains.
• Prompt-Based Editing: Modify an existing clip based on a written description without regenerating the entire video.
• Reference-Guided Generation: Use an input image to anchor style or character identity across multiple clips, rather than animating the image itself.
• Native Audio: Generate synchronized dialogue, ambient sound, effects, and music in the same pass as the video.

The API limitation is crucial for developers. If you are building a production pipeline via the developer API and expecting to feed it text prompts, you need to verify the specific model endpoint before you architect anything around it.

What Does Grok Imagine 1.5 Cost Across Different Plans?

Pricing varies significantly depending on how you access the model:

• Free Tier on grok.com: 5 credits per day, enough to run a few tests and understand what the model does, but not sufficient for consistent production work.
• SuperGrok Lite at $10 per month: Image and video generation at 480p resolution, clips up to 6 seconds, one AI agent, and longer chat windows than the free tier. A good starting point for creators wanting to try Grok Imagine seriously without a larger commitment.
• SuperGrok at $30 per month: Full Grok Imagine access with 720p output, up to 15-second clips, daily video render allocation, and unlimited image generation. This is the right plan for creators generating video regularly.
• X Premium+ at $40 per month: Higher throughput inside the X platform, priority routing, and ad-free X. Grok Imagine access is comparable to SuperGrok, worth it if you also want the platform benefits.
• xAI API with Pay-Per-Second Billing: 480p costs $0.08 per second; 720p costs $0.14 per second. Each input image adds $0.01. A 10-second 720p clip works out to $1.41. Running 100 clips costs $141 before any other infrastructure costs. The API is right for developers running automated pipelines at volume, but for solo creators generating a few videos a week, the per-second billing gets expensive fast.

How Should You Structure Prompts for Best Results?

Aurora renders each frame sequentially, first to last, which means actions you describe early in a prompt appear early in the video. Actions buried at the end of a prompt may not appear at all because the relevant frames have already been generated by the time the model processes them. This is the most common frustration creators hit in their first session.

Effective prompts need to be front-loaded and structured like a shooting script rather than a film synopsis. The recommended structure is: opening state first (what does Frame 1 look like, including camera position, subject position, light source, and atmosphere), then the progression (what changes over the clip's duration, described in the order it should appear), and finally audio and atmosphere (dialogue, ambient sound, and score belong at the end because they layer across the full clip rather than tying to a specific moment).

A well-structured prompt example: "Slow cinematic push-in as embers drift across the battlefield and the helmet's crest stirs in the wind." Notice the structure: camera move first, then subject behavior, then fine detail. Most people go wrong by writing prompts like a film synopsis, which gives Aurora too much to sort and the output tends to collapse into one static composition with motion happening around the edges rather than through it.

What Does the Rapid Rise of Grok Imagine Actually Signal?

The speed of xAI's ascent in video generation is remarkable. The company had no video product ten months ago. By January 2026, Grok Imagine debuted at number one on Artificial Analysis, beating Runway Gen-4.5, Sora 2 Pro, and Google Veo simultaneously. Version 1.5 extended that lead with a 52-point Elo gain. The Iliad trailer that Musk posted demonstrated that the model can sustain cinematic atmosphere across extended sequences with synchronized audio, a capability that represents a significant production ceiling for AI-generated video content.

For creators and developers evaluating whether this belongs in their workflow, the key takeaway is that the competitive landscape in AI video is moving extremely fast. Whatever limitations exist in the current version should be treated as a snapshot, not a stable verdict. Companies in this space are iterating at a pace that makes six-month-old benchmarks potentially outdated.