Elon Musk’s xAI launches Grok 3, challenges ChatGPT with new reasoning models and AI agent

Elon Musk led xAI has launched its Grok 3 series of models during a live-stream on Tuesday. The new Grok 3 model is said to have 10x more compute than the last generation and will compete with the likes of ChatGPT 4-o and Gemini 2 Pro. 

“Grok 3 across the board is in a league of its own. Even its little brother Grok 3-Mini is reaching the frontier across all the other competitors.” xAI team said during Grok 3 launch. 

Apart from the pre-trained model, xAI also launched a two reasoning model Grok 3 reasoning model (beta) and Grok 3 mini reasoning model. Notably, reasoning models are different from pre-trained models as they are meant to mimic human like thinking, meaning they take a bit more time to respond to a query but are also generally more useful for answering complex questions.

“We at xAI believe pre-trained model is not enough. That's not enough to build the best AI but the best AI needs to think like a human. To contemplate about all the possible solutions, self critique, verify all the solutions, backtrack and also think from the first principle.” xAI said while talking about need for building their reasoning model.

xAI said it has trained Grok 3 Mini reasoning model for a longer time frame than its elder sibling leading to it performing slightly better on some benchmarks. Needless to say, xAI claims that its new reasoning models perform better than OpenAI o1, o3 Mini (High), Gemini Flash thinking and DeepSeek R1 models in Maths, Science and coding benchmarks. 

xAI goes head to head with ChatGPT: 

xAI has also launched Deep Search which is the first generation of Grok AI agent that the company is marketing as a ‘next generation search engine’ that can help users ‘understand the universe’. 

“We're introducing a new product called Deep Search that is the first generation of our Grok agents that not just helps the engineers and research and scientists to do coding, but actually helps everyone answer questions that you have day to day.” xAI said in live-stream.

xAI says Deep Search analyzes the user intent, facts related to the query, reads to different sources and even cross validate the sources many times in order to make sure the answer it is providing is the one. 

Notably, OpenAI had recently launched its first AI agent, Deep Research which takes anywhere from 30 minutes to hours for getting answers to a complex query. The ChatGPT maker had then said that Deep Research can work independently to find, analyze, and synthesize hundreds of online sources and create a report at the level of a research analyst. 

While xAI did not give any benchmarks comparing Deep Search to OpenAI's Deep Research, one has to assume that the AI Agent battle has only just begun with Google, Claude and others likely to jump in soon.

How to use Grok 3 models? 

Grok 3 base model, reasoning models and Deep Search are being rolled out to the Grok website and app from today. However, the new series of models will intially only be available to Grok premium plus subscribers on X. Moreover, the company also announced a Super Grok subscription for Grok which it says will allow users to access the “most advanced capabilities and earliest access to new features”