Soon after the GPT-5 launch on Thursday, xAI CEO Elon Musk went on the offensive and stated that his company’s Grok 4 Heavy model still outperformed the new model running ChatGPT. Musk, who had unveiled the Grok series of models last month, also claimed that Grok 4 Heavy was the “most intelligent model in the world”.
“Bottom line though: Grok 4 Heavy was smarter 2 weeks ago than GPT-5 is now and G4H is already a lot better,” Musk wrote in a post on X (formerly Twitter).
The billionaire also shared that the Grok 4 Thinking model still ranked above GPT-5 High on the ARC-AGI 2 leaderboard — a benchmark designed to rigorously evaluate the abstract reasoning and problem-solving capabilities of AI models.
Musk further proclaimed that Grok’s latest language model update will arrive before the end of the year. At the same time, he announced that his company’s new Grok Imagine video and image generator would now be available to all users.
“Grok 5 will be out before the end of this year and it will be crushingly good,” Musk noted.
The xAI CEO also warned Microsoft chief Satya Nadella about OpenAI, saying that the ChatGPT maker would “eat Microsoft alive”.
ChatGPT gets a big upgrade:
Meanwhile, OpenAI, in its live-stream, stated that GPT-5 is a big upgrade over the previous model running ChatGPT in many areas including coding, math, writing, health, and visual perception. The company says its new model can generate apps, websites, and games in minutes with just natural language prompts, reducing the need for coding abilities.
GPT-5 is also the first-ever OpenAI model to have the in-built ability to choose a reasoning or efficiency model based on the query asked by the user.
The new AI model is being made available for free but with a usage limit, and once that limit is reached, users will automatically be shifted to a mini version of the GPT-5 model. OpenAI’s Plus users get access to a higher limit of GPT-5, while Pro users have access to the more powerful GPT-5 Pro model.
