Elon Musk is the CEO of multiple major companies, including Tesla, X, and the AI venture xAI, among others. On Wednesday, artificial intelligence company xAI launched its highly advanced flagship AI model, Grok 4.
Additionally, a high-tier AI subscription plan, SuperGrok Heavy, is charged at $300/month. Indeed, it's a significant move with a high price. Elon Musk’s new AI model is aimed at heavy developer use and is here to compete with the AI dominants, OpenAI’s ChatGPT and Google’s Gemini.
The latest model offers significant improvements in academic reasoning, multimodal, and coding capabilities. Additionally, it can analyze images and respond to questions like other AI models.
Musk stated during the live stream event of the Grok 4 launch that “The Grok 4 model has made major progress and can give answers to questions far better than PhD-level experts, nearly for all subjects. However, it has not yet developed new technologies or discovered new physics, but that wouldn’t take much time.” Musk even shared some of the limitations by adding, “It can make mistakes and lacks common sense reasoning.”
How is the New Model Performing on Benchmarks?
To demonstrate its significant improvements over previous versions, xAI tested Grok 4 as the frontrunner in most standard benchmarks, including Humanity’s Last Exam. The AI test features a challenging set of questions on subjects such as science, Mathematics, and humanities. Grok 4 scored exceptionally well, with 25.4 percent, without relying on any other tools. It defended the competitors in this benchmark, with OpenAI’s O3 model scoring 21% and Google’s Gemini 2.5 Pro scoring 21.6%.
When Grok 4 Heavy underwent the test on tools, it outperformed at the next level, achieving a 44.4% score while keeping Gemini 2.5 Pro behind with a 26.9% score using tools. Further, in the next ARC-AGI-2 test, Grok achieved top-class scores. The ARC-AGI-2 test is a benchmark that examines visual pattern recognition. Here, Grook 4 scored 16.2%, which was double that of its dynamic competitor, Claude Opus 4.
Stronger restrictions came about due to criticism of earlier Grok versions for their political and racial content. Musk promised to be more cautious and acknowledged his previous errors.
Staying Ahead in the Competition!
Grok is establishing itself as a significant player in the competitive AI market due to xAI's cutting-edge performance, rapid rollout, and premium pricing. With early access to Grok 4 Heavy and future releases, including an AI coding model (August), a multimodal agent (September), and a video-generating model (October), the SuperGrok Heavy package is designed for businesses.
XAI is launching Grok 4 with its API to help developers build applications with the new model. Although xAI's enterprise sector is still in its early stages, the company is looking to partner with hyperscalers to make Grok accessible via its cloud platforms.
xAI appears to have established a substantial foothold in today’s high-stakes industry; however, past controversies remain a point of complication in its expansion. Even though businesses are ready to put Grok first on their AI list, the downsides are still present. However, the AI race is nothing short of buzzing in!
Keep in the loop with all the trending news with us!
Recommended For You:
Understanding Multimodal AI: How Does it Work, and What are its Benefits?
Gen AI Outlook 2025: Key Trends Shaping Its Development and Future
