Better than GPT-4o: Grok-2 is now twice as fast!

xAI has provided a significant update to the Grok-2 and Grok-2 Mini it recently launched. It was stated that the AI ​​models experienced a two-fold increase in performance thanks to the rewriting of their codes. This shows that they can generate much faster answers.

According to a post by #xAI researcher Igor Babuschkin, Grok-2 Mini is twice as fast as it was two days ago. Babuschkin stated that all of the models’ inferences were rewritten using SGLang, an open-source system.

In addition to the performance increase, the researcher also stated that the models can give more accurate answers thanks to the new code. Thanks to this, $XAI can host the Grok-2 model on multiple servers and test it with more users.

The performance gains also made a difference in the leaderboard of the Lmsys Chatbot Arena, which tests artificial intelligence agents. Accordingly, Grok-2; came in second place in the overall ranking with 1293 points. It was neck and neck with the Google Gemini model, while it left OpenAI's #ChatGPT 4o model behind.

Grok-2 Mini, on the other hand, came in fifth place with 1268 points, leaving models such as GPT-4o Mini behind. Grok-2 is particularly leading in mathematical tasks, according to Arena. The high rankings of both models reflect areas such as coding, instructions, and reasoning.

The artificial intelligence model that allows you to create visuals with new image creation features and perform your daily tasks with #Grok is only accessible with Premium. With its ChatGPT-like theme, you can switch between Grok-2 versions at the top, and if you want, you can also switch to Fun Mode.