Elon Musk's Xai startup has announced the latest version of its flagship foundation artificial intelligence (AI) model, the Grok 4.
in Live stream In X, both masks boasted about the model still AI is simultaneously worried about its impact on humanity turn evil.
“This is the smartest AI in the world,” said Musk, surrounded by members of the Xai team. “In a way, it's scary.”
He compared the Grok 4 to “Super-Genius.” child“Just the “right value” of a sense of truth and honor must be instilled, so society must make profit from its progress.
Mask admits to being “worried,” saying, “It's a bit unsettling to create intelligence that is much greater than our own, is this bad or good for humanity?”
The Xai owner concluded, “Maybe that's a good thing.”
Musk said the Grok 4 is designed to perform at the “graduate level.” Many You cannot simultaneously talk about topics. It also allows you to process images, generate realistic visuals, and tackle complex analytical tasks.
Musk claims that Grok 4 will score perfectly on SAT and graduate level exams Like GRE flat Don't look at the questions in advance.
In addition to the model release, Xai has introduced SuperGrok Heavy, a subscription tier that costs $300 per month. Standard Grok 4 tier is available for $30 monthly, and The basic layers are free.
Openai, Google, Humanity and Confused We also announced a high-priced group. ChatGptPro, $200 a month. Gemini Ultra, $249.99 a month. Claude Max, $200 a month. And for $200 a month, Perplexity Max.
Please refer: Elon Musk Startup Xai launches apps that provide access to Grok Chatbot
Grok and X Turbulent Week
Grok 4 will be released with Grok 3. Linda Jaccarino resignationCEO of X.
Grok 4 Released in two standard configurations Grok 4 Premium “Heavy” version.
The heavy model features a multi-agent architecture that allows for collaborative inference on challenging problems.
Model Show progress In multimodal processing, Faster Inference and Upgrade User interface. According to Xai, Grok 4 can solve complex math problems, interpret images containing scientific visuals such as black hole collisions, and perform predictive analytics, such as estimating the team's chances of winning a championship.
Benchmark data shared by Xai shows that four heavy heavy models outperformed previous models in tests such as the last test of humanity.
Xai outlined an aggressive roadmap for the rest of 2025: boot Coding specific August AI, September multimodal agents, and fully-generated models video By October.
The release of Grok 4 has strengthened competition among major AI companies. Openai plans to deploy GPT-5 While Google continues later this summer Develop The Gemini series.
read more:
