Text-to-text? Done. Text-to-image? Mastered. Now, it's time for text-to-video.
The launch of OpenAI’s ChatGPT in 2022 sent shockwaves through the tech industry, triggering what many now refer to as the LLM (Large Language Model) arms race. Within weeks, every major tech giant scrambled to develop or acquire their own generative AI platform, attempting to match the momentum OpenAI had achieved seemingly overnight.
ChatGPT became a cultural and technological phenomenon. But amidst the excitement, controversy brewed—particularly around safety, censorship, and control over such powerful tools.
One of OpenAI’s earliest backers and co-founders, Elon Musk, grew disillusioned with the company's trajectory. Once a vocal supporter of OpenAI’s original nonprofit mission to democratize AI for the benefit of humanity, Musk became increasingly critical of what he saw as a pivot toward centralization and profit. He accused OpenAI of betraying its founding principles, becoming a “closed-source for-profit” entity too tightly intertwined with Microsoft.
In response, Musk launched xAI in 2023, which in turn created Grok AI family model.
With the LLM, Musk promised an unfiltered, witty, and rebellious counterpart to ChatGPT.
And this time, Musk’s xAI has now taken this strategy a step further with Grok Imagine.
Prompt:
A beautiful masked kunoichi ninja walking through the snowy streets of a dark, atmospheric Moscow night. She wears elegant black armor with flowing fabric accents, highlighted by gold jewelry and sharp eyeliner. Her mask is ornate with subtle patterns, revealing fierce,…— Elon Musk (@elonmusk) August 2, 2025
The AI-powered text-to-video tool now integrated into the X platform (formerly Twitter), allows users to generate short videos—up to six seconds or even up to six minutes, depending on implementation—complete with visuals and sound by simply typing descriptive prompts.
The underlying generation model powering Imagine is Aurora, xAI’s proprietary engine formerly applied to image generation.
Videos created through the tool are real-time editable, and static images can also be animated into dynamic visuals with audio. With this capability, Grok is positioning itself as a full-fledged creative assistant within X, offering tools for ideation all the way through to media generation.
Grok Imagine comes with a controversial twist: a "spicy mode" that reportedly permits nudity and more adult-oriented content.
While xAI hasn’t clarified moderation policies, this mode has raised alarms about deepfakes, nonconsensual content, and generative abuse, especially after prior controversies surrounding Grok’s AI companions and problematic output. Critics warn that such tools could be misused without robust safeguards.
Made with @Grok Imagine
pic.twitter.com/NXF5U0zD5k— Elon Musk (@elonmusk) August 3, 2025
Imagine @Grok
pic.twitter.com/1dkCHGAhjw— Elon Musk (@elonmusk) August 2, 2025
While tech reviewers and creators are excited about the imaginative potential of generating quick video stories, some worry about the ease of misuse.
Legislators and advocacy groups are sounding alarms over ethical implications, especially given AI’s ability to synthesize visuals that could be misrepresented or misappropriated. xAI has also faced backlash over earlier incidents of Grok producing antisemitic content and offering sexualized AI companions without robust filters.
Initially launched in beta, Grok Imagine is being rolled out to subscribers of the SuperGrok premium tier, which costs $30 per month. Interested users can request early access by updating the X app, navigating to the Grok section, and joining the waitlist.
A broader public release is expected around October 2025.
Grok Imagine is still early beta, so will improve almost every day.
A radical step-change in capability will come when we finish training our heavy video model on our 110k GB200s in a few months. https://t.co/VOPwF9LYgq— Elon Musk (@elonmusk) August 2, 2025
Made with Grok Imagine.
Have fun bringing art & sculpture to life just using your phone with @Grok.
pic.twitter.com/GSAXePs4Hd— Elon Musk (@elonmusk) August 3, 2025
Musk has framed this launch as a nostalgic return—“bringing back Vine, but in AI form.”
Whether users are creators dreaming of viral clips, a nostalgic Vine veteran, or an AI regulator wary of what’s next, Grok Imagine signals a major shift in how short-form content will be made—and who gets to make it. Written into its lines of code are both promise and controversy, as xAI embraces more expressive capability while negotiating culture, safety, and control in equal measure.
Imagine with @Grok pic.twitter.com/UIay5yNp97
— Elon Musk (@elonmusk) August 4, 2025
The launch of Grok Imagine followed the launch of Companions, which is a set of characters complete with their own personas that Grok users can chat with.
One particular character, named 'Ani,' speaks in a sultry tone and removes her clothing to reveal a lingerie set following a set of flirty interactions.
The character’s introduction was condemned by some users and organizations including the National Center on Sexual Exploitation, which advocates against pornography and sexual exploitation.
Musk stated following the launch that “customizable companions” would soon be available for users as well.
Animated with @Grok video.
We will allow you to iterate on audio tracks via an app update in a few days. pic.twitter.com/C7qRpCeaLm— Elon Musk (@elonmusk) August 2, 2025
Animated with @Grok Imagine pic.twitter.com/78AcUr4NsM
— Elon Musk (@elonmusk) August 3, 2025
As for Grok itself, the AI has evolved rapidly.
Since its debut in November 2023, Grok has undergone multiple transformative upgrades—most recently with Grok-4, launched in July 2025. This version introduces real-time tool integration, elevated reasoning abilities, and emotionally responsive speech companions designed to feel more human than ever.
Grok-3, released in February 2025, marked a major leap forward. Trained using ten times the compute of its predecessor—thanks to xAI’s custom-built Colossus supercomputer—it showcased remarkable improvements in scientific reasoning, logic, and advanced cognition.
Beyond the models themselves, the broader context of Grok Imagine reveals a massive infrastructure push by xAI. The company has reportedly deployed over 100,000 GPUs to power Colossus and support its AI systems. Its roadmap includes integration with real-time video chat, dynamic tool usage, and deeper system-level interaction directly through Grok.
Earlier milestones like Grok-2 introduced powerful text-to-image generation, while the original Grok was xAI’s first real entry into the competitive LLM arena.
Now, with Grok Imagine and its expanding multimodal capabilities, the platform is emerging as a legitimate challenger to OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude—especially in the realm of creativity, emotional intelligence, and real-world utility.
Grok Imagine is AI Vine!
Btw, we recently found the Vine video archive (thought it had been deleted) and are working on restoring user access, so you can post them if you want.— Elon Musk (@elonmusk) August 2, 2025