Grok didn’t hold back.

xAI has officially rolled out Grok 1.5 Vision, a major multimodal upgrade to its Grok AI assistant. This new model endows Grok with the ability to understand and process visual information, allowing users to upload images, documents, and screenshots for analysis. In line with its established brand, Grok's responses are infused with sarcasm and a lack of filter, offering a distinct alternative to the more guarded tones of mainstream AI chatbots.

Early adopters and beta testers have showcased the model's capabilities across social media, demonstrating tasks like identifying objects in a room, explaining complex charts, translating text from a photo, and deconstructing internet memes with humorous commentary. The release positions Grok as a direct competitor in the increasingly crowded multimodal AI space, challenging OpenAI's GPT-4V, Google's Gemini, and others, but with a unique, personality-driven angle that appeals to users seeking less sanitized interactions.

The launch follows a period of limited beta testing and is currently available to Premium+ subscribers on the X platform. This integration deepens the connection between Musk's AI ambitions and his social media platform, creating a closed-loop ecosystem for AI-powered content and interaction. The model's performance on standard benchmarks like MMMU and RealWorldQA has been published, showing competitive results, but its viral appeal lies primarily in its unconventional and often humorous output style.