Elon Musk To Open-Source Grok 2 Next Week Following OpenAI’s Release Of Open-Weights Models

OpenAI finally launching its first open models since GPT-2 is already having domino effects in the AI space.

After OpenAI had released two open-weights models yesterday, Elon Musk has said that xAI will open-source Grok 2 next week. Grok 2 was released in August 2024, and was one of xAI’s earliest models. xAI has since released Grok 3 and more recently Grok 4, which climbed to the top of most benchmarks.

Musk’s announcement about the open-sourcing of Grok 2 came after a user on X reminded him of a post he’d made in October last year. “Worth noting that xAI has been and will open source its models, including weights and everything. As we create the next version, we open source the prior version, as we did with Grok 1 when Grok 2 was released,” Musk had then said. An X user surfaced this post, and asked Musk, “Grok 2 and Grok 3 open-source when?”.

Musk saw the post and replied. “It’s high time we open sourced Grok 2. Will make it happen next week. We’ve just been fighting fires and burning the 4am oil nonstop for a while now,” he said.

xAI had open-sourced Grok 1, its very first model. But since then, it had released Grok 2, Grok 3 and Grok 4, and hadn’t open-sourced any of them. Musk now says that they’ll open-source Grok 2 next week. Musk has previously said that as they released a new version, they’d open-source the previous ones, so it’s likely that Grok 3 could also end up being released as open-source in the coming months.

This means that several US labs now have released open versions of some capable models. OpenAI has released two open-weights models, a 20b version and a 120b version. Meta has been releasing open models for a while now with its Llama series, and xAI will soon release Grok 2 as open-source. This leaves only Google and Anthropic among major labs which haven’t open-sourced any models of note, and it’s likely that given how other labs are open-sourcing their models, they too will follow suit. While the US had been vacillating on open-source, China has taken a sizable lead in the space, with models like DeepSeek and Kimi seeing mass adoption. The US, though, seems to be back in the game, and could soon end up competing with Chinese models for supremacy in the open-source space.

Posted in AI