Elon Musk’s company xAI has launched the beta version of Grok-2 and Grok-2 mini. The neural network can generate images in X and is said to have “advanced capabilities in chat, programming, and reasoning.”
— xAI (@xai) August 14, 2024
The advanced Grok-2 model and its lighter version are available to Premium and Premium+ users on the X platform. Developer access will also be provided through a corporate API by the end of the month.
Grok-2 outperformed Claude and GPT-4 in tests of the popular LMSYS benchmark, trailing only GPT-4o and Gemini 1.5 Pro.
“Grok-2 mini is our small yet functional model that offers a balance between speed and response quality,” the company blog states.
In a series of academic tests, which include reasoning, reading comprehension, mathematics, natural sciences, and programming, Grok-2 and Grok-2 mini demonstrated “significant improvements” over the previous model.
X users have shared images generated by the second version of Grok. It appears the AI does not have politically related restrictions.
Grok 2.0 …. Ohh boyyyy ??? pic.twitter.com/TjzB7WMhVp
— Benjamin De Kraker ?☠️ (@BenjaminDEKR) August 14, 2024
In one example, Grok 2.0 generated an image based on a request that ChatGPT deemed unacceptable.
Grok 2.0 will do political illustrations and real people, while ChatGPT refuses.
This instantly makes Grok 10x more fun…… pic.twitter.com/yDBJO0jWba
— Benjamin De Kraker ?☠️ (@BenjaminDEKR) August 14, 2024
To enable image generation, xAI collaborates with Black Forest Labs, “conducting experiments with their FLUX.1 model.”
xAI plans to integrate Grok-2 and Grok-2 mini into X’s AI features such as an enhanced search system and post analytics, and to release a preliminary version of multimodal understanding.
Back in April, xAI announced an updated version of the large language model Grok-1.5.
In the same month, the company introduced a new AI model capable of processing requests in various formats.