Elon Musk has revealed that xAI’s newest artificial intelligence model, Grok 4.5, has entered private beta testing at both SpaceX and Tesla. This marks the first confirmed deployment of the model within Musk’s companies ahead of a broader public release.
Grok 4.5 is built on xAI’s 1.5-trillion-parameter V9 foundation model, a significant leap from the previous 0.5 trillion-parameter "v8-small" version used on X, Musk’s social media platform. The model has also received supplemental training using data from Cursor, an AI coding assistant Musk’s SpaceX acquired earlier this year in a substantial deal.
Early evaluations indicate that Grok 4.5’s performance rivals and may even surpass Anthropic’s Claude Opus, a leading AI model. Musk also highlighted ongoing improvements through reinforcement learning and praised xAI’s internal coding framework known as the "Grok Build" harness for enhancing development efficiency.
In outlining the company’s forward plans, Musk said xAI aims to release new AI models monthly via SpaceX for the rest of the year. This points to an accelerated cycle of innovation and deployment within Musk’s technology ecosystem.
The V9 model’s training concluded recently, with the addition of Cursor-derived data playing a key role in refining Grok’s coding abilities. This integration represents a strategic push to advance AI tools capable of complex programming tasks. The announcement of Grok 4.5’s private testing phase underscores xAI’s rapid progression from research to practical application.