Grok users, both subscribers and non-subscribers, will enjoy unlimited access to the new Grok 4 model for a limited duration, as announced by xAI on August 10. This initiative is designed to maintain competitiveness as other AI firms introduce more sophisticated models and generative AI innovations.

“Just activate Auto mode, and Grok will direct complex inquiries to Grok 4. Prefer more control? Switch to ‘Expert’ whenever to consistently utilize Grok 4,” xAI mentioned in a statement on X.

This Tweet is currently unavailable. It might be loading or has been removed.

As xAI aims for a broader user community, the company may be trying to lure unsatisfied ChatGPT users following the introduction of OpenAI’s GPT-5. Many ChatGPT users on platforms such as Reddit have voiced that GPT-5 appears less human-like and more guarded in its responses. In light of this backlash, OpenAI’s CEO Sam Altman stated the firm will reintroduce its former model, GPT-4o, enabling other AI developers to provide alternatives with fewer restrictions.

Initial assessments of Grok 4, which debuted on July 9, have shown it to be remarkably advanced, outpacing rivals like DeepSeek and Claude. Nonetheless, the absence of safety measures raises possible concerns for users. A Mashable report uncovered that xAI’s new generative AI imaging tool, Grok Imagine, did not fulfill fundamental safety criteria, permitting users to generate sexual deepfakes.

Instructions for accessing Grok 4 for free

You can start utilizing Grok 4 immediately in the Grok iOS and Android applications, or through the X website.

In auto mode, the chatbot will assess if a request needs Grok 4’s comprehensive replies or Grok 3’s rapid feedback. Users can also opt for the “Expert” setting to obtain responses from Grok 4.

The organization intends to broaden its usage limits to accommodate the new model. More information on free tiers for Grok 4 has not yet been disclosed.

Disclosure: Ziff Davis, Mashable’s parent organization, initiated legal action against OpenAI in April, claiming it violated Ziff Davis copyrights in the training and functioning of its AI systems.