Elon Musk‘s company, xAI, has introduced two new models of its AI chatbot, Grok, which can now generate images and has advanced reasoning capabilities, such as writing code. The models, Grok-2 and Grok-2 Mini, are currently in beta testing and available to users subscribed to the Premium and Premium+ packages on Musk’s social media platform, X.
In a blog post on August 14, xAI announced the release of Grok-2, describing it as a major upgrade from the previous Grok-1.5 model, with enhanced features in chat, coding, and reasoning. The company also introduced Grok-2 Mini, a smaller but still capable version of Grok-2. An early version of Grok-2 was tested on the LMSYS leaderboard under the name ‘sus-column-r.’
The image-generation capabilities of Grok-2 have generated significant attention. While some beta testers and app researchers noted improvements in the chatbot’s ability to write code, generate text, and summarize news, the most talked-about feature is its ability to create images.
Following the limited release, users started sharing images on X that were reportedly created by Grok. Concerns were raised about the lack of safeguards in the text-to-image model, as it could generate controversial, photo-realistic images of politicians and other public figures.
With the upcoming US elections, there are growing concerns about the potential misuse of AI-generated content to spread misinformation on platforms like X. Some of the AI-generated images suggest that Grok may be integrated with FLUX.1, a new diffusion model developed by Black Forest Labs, an AI startup founded by former engineers from Stability AI.
xAI plans to make the new versions of Grok available to developers through an API in the coming weeks. The company is also looking to roll out Grok-2 and Grok-2 Mini on X to enhance search capabilities, post analytics, and reply functions. Additionally, Grok’s multimodal understanding capabilities may become accessible to both X users and developers.
