xAI makes Grok 2.5 open source and plans the same for Grok 3

Many of the concerns about artificial intelligence can be overcome through transparency. And it is in the name of transparency that Elon Musk has announced the open sourcing of its Grok 2.5 model.

More than this, the AI firm will also make the Grok 3 model open source is around six months.

The owner of X and xAI said simply: “The @xAI  Grok 2.5 model, which was our best model last year, is now open source. Grok 3 will be made open source in about 6 months”.

Effective immediately, the source code is available for anyone to see, download and use via Hugging Face. The repo says of Grok 2:

This repository contains the weights of Grok 2, a model trained and used at xAI in 2024.

Musk made the announcement in a brief, low-key post on his own social platform X:

Grok to go

While the news will be welcomed by anyone looking to delve into the inner workings of Grok 2.5, Musk and xAI are hardly leading the way here. Meta has already made its own Llama model available on an open source basis.OpenAI, like Meta, has also released its AI models on an open source basis, and there is no indication that this is not something that will continue.

It seems that Musk is working to a yearly cycle. With Grok 2.5 having been xAI’s model from 2024, it is being open sourced now, in 2025, Grok 3 launched in February this year, and the six month timeline would take us up to around February next year.

Musk has previously said that Grok 5 is expected to launch by the end of 2025, so – all things being equal – we can probably expect to see that particular model open sourced in late 2026 or early 2027.

There have been various controversies surrounding Grok, including giving racist answers to queries, and user data being leaked.

This is just the latest example of big companies understanding, and publicly demonstrating, the importance of open source software. It is something that even the likes of Microsoft has been engaged in, and there is a lot of sense in making source code publicly available and reusable. There is much to be learned from looking at source code – it helps to determine how a project operates, but it also gives more eyes the opportunity to spot errors and come up with ideas for how things could be improved.

Image credit: Rafael Henrique / Dreamstime.com