xAI has done an amazing job playing catch up to competitors and they have just dropped a SOTA model that outcompetes other billion dollar companies in the same space.

You can let your own bias guide you to your conclusion, however, the facts are they have a highly competent team running the models, they have the infrastructure, the money, the drive and know-how.

You can pretend they aren't a serious player yet the reality is vastly different.

Elon Musk cofounded and funded OpenAI.

I use Grok, ChatGPT, and Gemini. They are all excellent, state of the art, and have their unique strengths and weaknesses.

xAI is an attempt by Elon to remain relevant and have a "woke" model that isn't moralizing him when he asks racist questions

OpenAI is Altman's attempt to use brand perception to con everyone into thinking they aren't loosing the lead on the field they pioneered, while hyping up investors that AGI is around the corner. And except for the hunt for AGI they have given up everything they originally stood for, leading to the mocking term ClosedAI

Llama would not be noteworthy if not for the fact that it's open weights

Gemini had an embarrassingly terrible start considering the amount of data and AI talent Google has at its disposal. Their recent models are pretty good, but their bad start combined with the cheap models they roll out to a wide consumer base still hurt their perception. Google's models are probably the first thing people think of when talking about bad AI

DeepSeek and Qwen are impressive but Chinese

You can find reasons for all of them why they are embarrassing places to work at. Yet people do work there. And judging from the results (both Grok3 and Grok4) xAI seems to do just fine on training data and attracting talent