xAI, the artificial intelligence company led by Tesla CEO Elon Musk, has expanded its API with new image generation capabilities. The API, launched in October 2024, now features Grok-2-Image-1212, a model capable of generating up to 10 images per prompt at a cost of $0.07 per image in JPG format.
However, xAI has acknowledged the limited functionality of its current image generation capabilities. Customization options such as adjusting quality, size, or style are unavailable, and the model lacks the extensive features seen in competitors like OpenAI’s Sora or MidJourney.
Additionally, xAI’s API incorporates a chat model that revises user prompts before submission. This pre-processing step aims to improve the prompt’s accuracy for better image generation results.
xAI’s Growth Strategy
The introduction of image generation is part of xAI’s expansion strategy. The company is currently exploring further investments in its Memphis data center, a key site for AI model training. Furthermore, xAI recently acquired HotShot, a video generation and text-to-video startup, suggesting ambitions to compete in the AI-powered media generation space.
In parallel, xAI launched Grok 3 last month, positioning it against OpenAI’s GPT-4o and DeepSeek’s R1 model. Musk claimed that Grok 3 possesses 10 times the computing power of its predecessor and benefits from a broader dataset that includes materials like court case filings.
“Grok 3 is an order of magnitude more capable than Grok 2,” Musk said at the launch. “[It is a] maximally truth-seeking AI, even if that truth is sometimes at odds with what is politically correct.”
Controversies Surrounding Image Generation
Despite its advancements, xAI’s image generation history is marred by controversy. Grok 2 came under fire for its lack of content regulation. The model generated offensive and graphic images, including depictions of Donald Trump piloting an aircraft towards the World Trade Center and Ronald McDonald holding an assault rifle.
These images, described as reckless and irresponsible by civil rights attorney Alejandra Caraballo, underscored the risks of AI misuse.
“Grok has absolutely no filters for its image generation,” Caraballo remarked. “This is one of the most reckless AI implementations I’ve ever seen.”
While xAI has not commented on whether it has implemented stricter safeguards for its new image generation model, the industry will undoubtedly scrutinize its future outputs.
Competitive Landscape
With its limited customization options and controversial track record, xAI’s offering faces stiff competition. OpenAI’s Sora, for example, allows users to create highly detailed, customizable videos with enhanced safety mechanisms. Similarly, platforms like MidJourney and DALL·E 3 continue to push the boundaries of responsible AI image generation.
However, xAI’s recent acquisitions and investments signal its intent to close the gap. The introduction of text-to-video capabilities through HotShot may further enhance xAI’s standing in the AI-generated media landscape.
