Elon Musk’s Grok 2 made waves when it was released last week to the public and its available through an $8 USD monthly subscription on X (formerly twitter).
Grok-2 comes in two versions, Grok-2 and Grok-2 mini. Grok2 mini is specifically designed to be less powerful but faster. Speed of both versions has now been increased. They can now analyze information and give output much faster after two developers at xAI completely rewrote the code stack in the last three days.
As xAI developer Igor Babuschkin posted this afternoon on the social network X under his handle @ibab:
“Grok 2 mini is now 2x faster than it was yesterday. In the last three days @lm_zheng and @MalekiSaeed rewrote our inference stack from scratch using SGLang. This has also allowed us to serve the big Grok 2 model, which requires multi-host inference, at a reasonable speed. Both models didn’t just get faster, but also slightly more accurate. Stay tuned for further speed improvements!”
Now according to Babuschkin’s post, the two genius developers who rewrote the code in three days to make Grok 2 faster are are Lianmin Zheng and Saeed Maleki. Apparently, they relied on SGLang, an open-source highly efficient system for executing complex language model programs.
Babuschkin wrote in one of the tweet’s replies on X that the main reason of using Grok-2-mini over the full Grok-2 model is its enhanced speed.
Babuschkin further emphasized that xAI will further enhance the processing speed of Grok-2-mini, which will make it more desirable for users who are seeking high performance with less computational overhead.
The success of these models bear witness to xAI’s innovative desire to get the maximum of AI’s full potential. As xAI continues to refine its models with time, we can expect even greater advancements in speed and accuracy of both Grok-2 Grok-2-mini, keeping them at the helm of AI development.
This post was first created by me on Medium