Elon Musk’s XAI launched Grocke 3, challenged Chatgate with new Reasoning Model and AI agent

Elon Musk’s LED XAI launched its Grocke 3 series models during a live-stream on Tuesday. The new Groke 3 models are said to have 10x more calculations than the previous generation and this chat will compete with the choice of 4-o and Gemini 2 Pro.

“Grocke 3 across the board is in one of its own leagues. Even its younger brother Grocke 3-Min is reaching the border among all other contestants. The XAI team said during the Grock 3 launch.

In addition to the pre-eminent model, XAI also launched a two Reasoning Model Grocke 3 Reasoning Model (Beta) and Grocke 3 Mini Reasoning Model. In particular, logic models differ from pre-influential models because they are like thinking to mimic human beings like thinking, which means they take a little longer to respond to a querry, but usually Are more useful for answering complex questions.

“We believe in XAI that pre-educated model is not enough. This is not enough to make the best AI, but the best AI needs to think like a human. To contemplate all possible solutions, self -critical, verify all the solutions, box and think of the first theory. “XAI said talking about the need for the construction of his argument model.

XAI said that it has trained the Grok 3 mini region model for longer than her elder brother -in -law, which is performing a little better at the benchmark. Needless to say, XAI claims that its new logic models perform better than the Dipsec R1 model in Openai O1, O3 Mini (High), Gemini Flash Thinking and maths, science and coding benchmarks.

Xai goes to the head with chat:

XAI has also launched a deep discovery that is the first generation of the Groke AI agent that the company is marketing as the ‘next generation search engine’ that can ‘help users understand the universe’.

“We are introducing a new product called Deep Search that is the first generation of our grooc agents that not only helps engineers and research and scientists coding, but in fact help everyone’s questions answer questions What you have day by day. ” XAI said in live-stream.

XAI says that deep search analyzes the user’s intentions, the facts related to the query, read different sources and even validate the sources many times to ensure that it can be ensured that the answer provided by this He is doing one.

In particular, Openai recently launched its first AI agent, deep research that takes anywhere from 30 minutes to hours to get answers to a complex query. The CHATGPT manufacturer then stated that intensive research can work freely 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 benchmark compared to deep discovery for deep research by OpenaiI, one would have to believe that the AI ​​agent’s fight has only started with Google, Cloud and others, who likely to jump soon Is.

How to use Grok 3 model?

Grocke 3 base model, Reasoning model and deep search is being rolled out for the Groke website and app from today. However, the new series of models will be available only to the Groke Premium Plus customers at X only. In addition, the company also announced a super grouke subscription for the grouke, which states that it would “allow users to reach the most advanced capabilities and new as soon as possible.