Over the weekend, Elon Musk’s artificial intelligence company XAI released Grok Imagine, a new generation AI tool for generating images and videos. Grok Imagine is now available to pay XAI subscribers in Grok iOS and Android apps.
Musk has been hyping the project on X, sharing photos and videos of Grok users. This includes some mild NSFW content, the Grok app tagged “spicy”.
This tweet is currently unavailable. It may be loading or being deleted.
AI video is exciting and frankly frightening new areas of the AI industry. For supporters, this technology provides artists with a new medium of creativity and can reduce the cost of animation and film production. For critics, AI video poses serious risks to the deep sexual risk and misinformation.
With this debate on hold for the time being, I want to see how Grok imagines it compared to Xai’s biggest competitor. As I’ve written before, Google’s WEO 3 AI video model currently comes with unexpected lifespan videos. Then there is Sora, from Chatgpt-Maker Openai. In addition, the popular AI Image Generator Midjourney recently launched its own AI video generator tool.
So, how does Groke imagine comparing to his competitors? To be honest, I was not impressed.
Yes, Grok Imagine is brand new, and Musk recently said on X, “It should be better every day.” However, at the time of writing, it seems to be lagging far behind its competitors.
Let me show my work.
Comparing Grok Imagine AI videos with matches
Mashable recently wrote about the trend of viral AI videos – Animals’ security camera lens jumps onto a trampoline and engages in similar antics. So I used a simple tip to test Grok Imagine, VEO 3, Sora and Midjourney: “The security camera lens of a rabbit who jumped onto a trampoline at night.” It’s simple, right?
First, I should point out that there is a big difference between Veo 3 and Grok Imagine. Google’s VEO 3 model can generate videos based on text prompts. Just describe the video you want and the rest will complete the rest. However, tools like Midjourney and Grok imagine a generation that only provides text to images. After an image is generated or uploaded, the user can animate it and convert it into a short video clip. In this sense, Grok Imagine is already on the back foot compared to OpenAI and Google.
Mixable light speed
With these warnings, let’s study the results, and I also shared the results on X.
I put the test prompts into my Grok and returned these disappointing images.
Credit: Screenshot courtesy of Grok/Timothy Beck Werth
Credit: Screenshot courtesy of Grok/Timothy Beck Werth
I selected the worst of these images and created this short video:
This tweet is currently unavailable. It may be loading or being deleted.
…Okay? Intermediate or Umas the children say.
But this also suffers compared to other AI video tools.
As shown in the video, Google VEO 3 and Sora performed better in the same prompt:
This tweet is currently unavailable. It may be loading or being deleted.
Finally, Midjourney animates XAI-like images and despite trying twice, he was able to make better images and videos. It produces images and videos with a granular appearance of the surveillance lens.
Images generated by AI.
Credits: Timothy Baker Worth/Juni
This tweet is currently unavailable. It may be loading or being deleted.
Audio is also a major drawback of Grok Imagine. Although VEO 3 can synchronize with video to produce sound effects and coherent conversations, the audio imaginary videos I found on Grok are limited to rough sound effects and gibberish.
Musk compared Grok Imagine with the modern grape application and wrote on X: “Grok Imagine is optimized for the most interesting and shareable content.”
In my initial test, Grok Imagine seems to optimize creating two types of images and videos: memes and anime. If you want to animate memes or create a sex hint video of anime girls – then I think Grok Imagant will do that. But other than that, I can’t say I’ve left a deep impression on me.
In one area, Grok Imagine does shine, and that’s the speed. So far, I’ve found that it produces images and videos much faster than its competitors.
This tweet is currently unavailable. It may be loading or being deleted.
Mashable is exposed to XAI and we will update this story if we receive a reply.
Disclosure: Mashable’s parent company Ziff Davis filed a lawsuit against Openai in April, accusing Ziff Davis of infringing on Ziff Davis’ copyright in training and operating its AI system.
